Review: Opportunities for data-based on-farm estimation of greenhouse gas emissions from ruminant farming systems.

Livestock farming has become increasingly digitalised in large parts of the world. Sensor technologies can be used to mitigate emissions from farming in several ways, including improved efficiency of operations. The data that farmers collect to monitor and optimise their production offer a possibility to increase the accuracy of greenhouse gas (GHG) emission estimation if utilised fully. However, technologies developed for farm management may not meet requirements for monitoring of GHG emissions where high accuracy is required. The cost-benefit and reliability of farm-level data collection need to be considered before investing effort in their widespread adoption. The lack of access to data and poor interoperability of the data across different software systems also currently limits the use of data; however, there are several data sharing initiatives, such as data spaces, that aim to develop the farmers' right to data sovereignty and provide contractual and technical solutions for increasing data availability. This review assesses the state-of-the-art of digital data collection and sharing usable for GHG accounting from a ruminant farming systems perspective, with a focus on commercially available on-farm technology. A review of the accuracy of different methods and two case studies on the impact of the uncertainty of digital data sources on CH4 emission prediction uncertainty quantified using Monte Carlo simulation are presented. In conclusion, the greatest potential for automating the life cycle assessment process and coverage of production for national inventories exists for dairy cattle, sheep and goat farms with daily milk yield recording and high technology feeding systems and the lowest for pasture-based meat production systems. The use of on-farm data for GHG accounting provides an opportunity for more targeted mitigation and even differential product pricing favouring the lowest emitting farms within a region. The reliability as well as contribution to overall emissions of each additional data source collected from farms should be assessed before it is taken into use, and calibration protocols for farmers to ensure accurate data should be developed and their application monitored.