Welsh food wholesaler Castell Howell Foods received only 41% of the vegetable volumes its growers had forecasted, exposing a critical gap in supply chain planning. Data Dynamics is building an AI system that fuses satellite imagery, phenological models, and historical yield data to predict crop harvests months in advance across key Welsh growing sites. The problem is that unpredictable harvests force wholesalers to scramble for last-minute supplies, driving up costs and waste, while small farms—which make up 77% of Welsh agricultural businesses—bear the brunt of unstable pricing. If the system works, Castell Howell can shift from reactive procurement to coordinated planning, locking in supply and price stability before the growing season ends. The pilot focuses on two sites, but the technology is designed to scale across the Welsh food sector, which contributes over £2.23 billion in GVA to the economy. This is applied research with a direct commercial partner, not fundamental science—success means a working forecasting tool, not a new theory.
View original technical description
Data Dynamics proposes to develop an AI-driven crop yield forecasting system in partnership with Castell Howell Foods, Wales's leading independent food wholesaler. This innovative system will enhance supply chain resilience by providing accurate yield predictions for key vegetable and arable crops across Castell Howell's network of Welsh growers and suppliers. The Welsh food and drink sector contributes over £2.23 billion in GVA to the Welsh economy, with the food foundation sector being particularly vital to rural communities in Mid and North Wales. However, agricultural supply chains face increasing challenges in managing production variability and maintaining consistent supply to meet demand, as demonstrated by Castell Howell's recent experience where growers delivered only 41% of forecasted volumes. Our solution integrates multiple data streams to deliver comprehensive yield forecasts throughout the growing season. High-resolution satellite imagery will monitor crop development patterns, while advanced phenological modelling will track growth stages and environmental responses. Machine learning algorithms will analyse this data alongside historical yields to generate accurate predictions for vegetables crucial to Castell Howell's supply chain. The initial pilot with two key growing sites will demonstrate how improved yield forecasting can transform supply chain management. By providing earlier and more accurate crop yield predictions, the system will enable Castell Howell to optimize its procurement planning, reduce waste, and better coordinate with their network of Welsh growers. This enhanced visibility will help stabilize supply and pricing, benefiting both producers and distributors. The project aligns with Welsh Government objectives for agricultural sustainability and rural community resilience. More accurate yield forecasting will help stabilize agricultural supply chains, particularly benefiting the small and medium-sized enterprises that comprise 77% of Welsh agricultural businesses. By bringing together expertise in remote sensing, artificial intelligence and agricultural science, this project will create an innovative and robust solution for strengthening Welsh agriculture. The focus on direct implementation with Castell Howell Foods ensures the technology will deliver immediate value while establishing a foundation for broader adoption across Welsh food supply chains.
Plain English summaries and category classifications on this site are generated by AI and may not perfectly reflect the original research.
Is something wrong? Let us know