DIVINE Pilot 2 - Case Study

DIVINE Pilot 2 Case Study - Crop Yield Prediction Model

Pilot Overview

This pilot is targeted at improved crop yield prediction in Ireland. The pilot solution is tested on 56 ha of cereal crops in Kildare, Ireland. Cereal production is important in Ireland with about 150,000ha of winter cereals grown and 180,000 of Spring crops.

The overall objective of the pilot is the delivery of an easy-to-use crop yield prediction model and decision support tools for production optimisation.

Target Stakeholders & Problems to be Addressed

This solution is targeted at arable farmers, agronomists, advisors and agri-tech solution providers. The pilot addresses inhibitors to effective digital support including manual data collection, lack of interoperability and connectivity between data sets and technologies and grower trust regarding data collection. Furthermore, variable weather, disease pressure, weed issues, and variety genetics can make yield predictions challenging.

Initially based on UCD college research farm and the linked research farms that the university uses for crop rotation purposes, typically this will cover circa 56 ha. Additionally, through open day events, it can then be demonstrated to tillage producer stakeholders as well as main actors in the supply chain e.g. seed/chemical suppliers, advisors, crop buyers etc.

Key Activities

A comprehensive scoping of private and public data sources has informed this pilot design and rollout. A number of internal and external data sets have been integrated including;

  • Meteorological data
  • Department of Agriculture, Food, and The Marine varietal test data
  • Earth observation data products
  • Irish Soil Information System
  • Sensor recordings
  • Production practice information from multiple sources (e.g. FMIS) including pesticide applications, seeding/ Harvesting, Yields

Decisions support tools developed and tested include;

  • Variety selection tool,
  • automated plant counting from imagery
  • cereal crop production blueprint

This pilot has used data from 2023, 2024 and 2025 to test and refine the developed solutions.

Outcomes & Impact

Cereal crop production is extremely complex with up to 200 individual decisions necessary at various stages from seed selection to harvesting. Each decision must consider several context-specific factors. The cereal crop production blueprint and its digitisation simplifies this process.

Farmers benefit from having a tool that will help them to predict their crop yield from real-time inputs, giving them the opportunity to implement best practice, reduce inputs where possible, and thus increase farm economic viability. The tools developed have the potential to optimise yield management throughout the growing season from establishment until the final spray programme thus enabling growers to adjust their production strategy based on in-season data collection.  This system also enables growers to estimate their yield potential from the germination results and hence tailer their management strategy.  Farmers get customised advise based on their priorities and specific situations. Decisions can quickly account from a huge variety of factors including current and future weather, varietal specificities, soil types, available establishment methods, soil moisture and temperature and incidence of local diseases and pests. This technology will benefit farmers and agronomists/ advisors and has the potential to attract new entrants to crops for example livestock farmers, young people, people with non-agricultural backgrounds. These services allow farmers to get better insights on the crop developments in their region, but it also allow agri-tech providers to enhance their existing advanced decision support systems with an additional valuable source of information.

We have presented this to lots of attendees both national and international. We have also presented to the Minister for State for Agriculture at an open day at Lyons earlier this year as well as to senior advisors in the Department of Agriculture. We would estimate that the attendees would be circa 80 people excluding the Department of Agriculture staff etc in the last two years.