Livestock manure is a major source of methane emissions. Direct emission measurement is time-consuming and expensive while mathematical models offer a cost-effective alternative to achieve emission estimates for specific source categories. However, their accuracy must be validated by measurements. In this study, the performance of an Arrhenius-type model and the Anaerobic Biodegradation Model (ABM) was investigated at a Danish growing-finishing pig farm over three batches. Continuous methane measurements were conducted using cavity ring-down spectroscopy, and slurry samples were collected to determine the specific methane production rate. Slurry temperature and chemical composition of slurry, faeces, urine, and feed were used as model inputs. In a fourth batch, slurry samples were collected before and after flushing to assess methanogen identity, abundance, and activity. The methane production rate from slurry samples was markedly lower than the in situ measured methane emissions. The ABM and Arrhenius model underestimated the emissions by 19 and 40%, respectively, although dynamics were captured relatively well and the estimation was improved after optimisation. The microbial analysis revealed that Methanosphaera was the most abundant methanogenic genus but exhibited only marginal activity. Methanoculleus was the most active genus, except in the late stage of the batch where Methanosarcina dominated, contributing over 90% of the methanogenic activity. While ABM shows promise for achieving farm-specific emission estimates, more studies at laboratory-scale are needed to optimise key parameters, and additional farm-specific data should preferably be identified in order to represent real-world variations in methane emissions. • A Danish pig farm served as case study to validate the ABM and Arrhenius model. • Methane emissions were underestimated by both models. • The ABM outperformed the Arrhenius model also after optimisation. • Residual slurry emerged as an important factor to predict methane emissions. • Methanogenic presence did not correspond to methanogenic activity.
García et al. (Tue,) studied this question.