Objectives of the service
Many growers still decide when to plant, fertilise and harvest from experience and the calendar, which is increasingly unreliable as weather becomes more variable. Pheno-AI addresses this by turning freely available satellite and climate data into clear, field-level guidance.
The service combines a heat-accumulation model (Growing Degree Days) with satellite vegetation observations to forecast key crop development stages and a likely yield range for each field, and it states how confident each forecast is. This activity covers three areas: engaging growers to capture their needs, assessing whether the forecasting approach is technically sound on free satellite data, and assessing whether the resulting service is commercially sustainable. The aim is a practical, low-cost advisory service that works even where ground data and connectivity are limited.
Users and their needs
The targeted users are medium and large crop growers, cooperatives, agribusinesses and agricultural advisors, together with adjacent users such as crop insurers and input suppliers. Users are engaged in Kenya (maize and beans) and at calibration sites in Portugal and Germany (cereals and permanent crops). Their main needs are:
- guidance on the right time to plant under variable rainfall;
- decision support on fertiliser timing and application rate;
- guidance on harvest timing and method;
- improved crop-forecasting accuracy and yield- and climate-risk management;
- access without on-farm hardware, usable in data-sparse, low-connectivity settings.
The targeted users are located in Kenya, Portugal and Germany.
Service/ system concept
For each field, the service delivers forecast emergence, flowering and harvest windows, a yield range with risk flags, and timing recommendations, through a web application and an interface for partner systems. It combines daily temperature data (used to compute heat accumulation) with satellite vegetation observations; a model fuses these into a forecast and an associated confidence range. Where cloud blocks the optical satellite view, radar satellite data and statistical methods keep the service running, with the confidence range widened accordingly. The approach is software-only and runs on cloud infrastructure, so it needs no equipment on the farm. The high-level architecture is shown below.
Space Added Value
The service is built on European Space Agency Copernicus Earth Observation data. Sentinel-2 provides free optical imagery used to track crop development, and Sentinel-1 provides free all-weather radar imagery used when clouds obscure the optical view. Combining these free satellite assets with open climate data makes field-level forecasting possible at very low cost and across large areas, including regions where ground sensors and weather stations are scarce. Ground-based or hardware-dependent methods from existing providers cannot achieve this coverage economically at the same scale, which is the core added value of using space assets here.
Current Status
The project has completed its Kick-Off and reached the Mid-Term Review. A champion customer in Kenya has signed a proof-of-interest and has helped open a pipeline of large-scale growers and a county agriculture ministry. User needs and user requirements have been captured and documented; the technical concept and system architecture have been defined on free Earth Observation and climate data; and the market, competitive landscape and business model have been analysed. Work in progress includes deepening user validation across Kenya, Portugal and Germany and preparing the technical-feasibility evidence for the Final Review.
