Classification of sheep urination events using accelerometers to aid improved measurements of livestock contributions to nitrous oxide emissions

Livestock emissions account for 74% of nitrous oxide contributions to greenhouse gases in the UK. However, it remains uncertain how much is directly attributable to localised sheep urination events, which could generate nitrous oxide emission `hot spots'. Currently, IPCC emission factors are mainly extrapolated from lowland grazing systems and do not incorporate temporal or spatial factors related to sheep behaviour and movement. Being able to gather data that reliably measures when, where, and how much sheep urinate is necessary for accurate calculations and, to inform best management practices for reducing greenhouse gas emissions and minimizing emission-based climate change. Animal-attached movement sensors have been shown to be effective in classifying different behaviours, albeit with varying classification accuracy depending on behaviour types. Previous studies have used accelerometers on cattle and sheep to assess active and non-active behaviours to help with grazing management, although no study has yet attempted to identify sheep urination events using this method. We attached tri-axial accelerometer sensor tags to thirty Welsh Mountain ewes for thirty days to assess if we could identify urination events. We used random forest models using different sliding mean windows to classify behaviours. Urination had a distinctive pattern and could be identified from accelerometer data, with a 5 s window providing the best recall and a 10 s window giving the best precision. 'State' behaviours considered (foraging, walking, running, standing and lying down) were also identified with high recall and precision. This demonstrates the extent to which the identification of discrete `event' behaviours may be sensitive to the window size used to calculate the summary statistics. The method shows promise for identifying urination in sheep and other livestock, being minimally invasive compared to other methods, and has clear potential to inform agricultural management practices and policies.