Soil Moisture Prediction
Challenges
- A leading agribusiness operating in water-stressed regions set out to increase yields while reducing irrigation costs.
- It needed better precision in water management, and stronger sustainability and resource efficiency.
Approach
An AI-powered virtual soil moisture sensor predicts soil moisture to roughly 1% error and issues field-level irrigation recommendations.
The system integrates weather, soil, crop and irrigation data into a three-step prediction model and delivers recommendations through a web application for field staff.
Key features
- High-precision soil moisture prediction from weather stations, sensors and crop data
- Automated irrigation recommendations per 50 × 50 m grid
- Integration with existing irrigation infrastructure, including pivot systems
- Real-time interface for farm operators with actionable recommendations
Impact
+1.2–3.5%
EBITDA uplift from higher yield and water savings
~1%
Prediction error on soil moisture
50 × 50 m
Grid resolution of irrigation recommendations