Long-day onion yield and storage forecasting model for growers

Overview

This solution is a yield and storage forecasting model designed specifically for long-day onion crops. It addresses a critical gap in the onion supply chain: the lack of systematic in-season crop metrics for yield, canopy development, and bulb size that would enable knowledge-based decision making. By providing growers with predictive insights on yield, the quality of produce entering storage, and expected shelf-life once stored, the model supports more efficient harvest planning, storage management, and waste reduction further up the supply chain.

Technical specifications

The model is built on NIAB's commercially deployed potato yield forecasting platform, which has been in commercial use since 2012 and received industry recognition in 2014. The adaptation for long-day onions uses three primary input categories: historic farm and variety data, crop canopy satellite imagery, and meteorological data. The platform also integrates historic onion storage data to make predictions on genetic storage potential and disease levels likely to be present on material entering storage.

Key features:

• Predicts yield and graded bulb size fractions with demonstrated correlation to actual trial results
• Estimates storage quality and shelf-life based on genetic and environmental factors
• Requires only minimal inputs to run: location, variety, and achieved population count
• Optional mid-season yield measurement can be incorporated through data assimilation to improve accuracy
• Reflects crop stress where it is measurable in the canopy

The model is limited to onion crops grown from seed and assumes that water, nutrient, pest, weed, and disease pressures are managed or absent except where stress is visible in canopy imagery.

Technology readiness level

Initial adaptation of the potato model for onion yield prediction, using data from research trials, has already shown good correlation to actual yields and graded bulb size fractions. The underlying potato platform is commercially proven and in active use by commercial growers. An initial working model for onions could be developed within 12 months from project start, with further refinements expected over multiple growing seasons. NIAB's 30-plus years of onion yield and storage trial data provides a strong foundation for model development, and access to additional historic commercial and breeding line data across diverse seasons, soil types, and storage conditions would further strengthen predictive accuracy.