The challenge

AgriWebb has historically focused on record collection of actuals entered by the producers. The Foragecaster initiative will use these best of breed models as inputs as well as machine learning to create a livestock supply and sustainability forecaster. Foragecaster will use the 60 million animals tracked over the last 8 years from over 12,000 producers in AgriWebb to develop machine learning models to help predict livestock growth. The producer will get a probabilistic forecast for pasture growth and availability as well as the sustainability metrics for their natural capital. These will be exposed to the producer through new planning tools including a grazing planner and a scenario planner. 

Project outputs

• Grazing Planner – the pasture and livestock forecasting models will feed into a new grazing planner to support decision-making by taking into account climate forecasting and growth models, allowing the producer to forward plan
• Climate information – localised historical and forcast weather information will be provided to the farmers
• Sustainability metrics – the Foragecaster predictive models will be used to determine sustainability metrics both current and also forecasting into the next 6 months. The sustainability metrics will include the amount of carbon sequestration through vegetation and soil and the amount of emissions from the livestock as well as other factors that contribute to natural capital such as biodiversity.

Partner contributions

A 6 month pilot started in January 2023 with the below partners for the researchers to understand the problem space and the existing data, hire the required postdocs and researchers, and to start prototyping different approaches.