Improving Wheat Rust Early Warning in Ethiopia

In Ethiopia’s wheat-growing regions, farmers live with a constant uncertainty. Wheat rust—a fast-moving fungal disease—can spread silently and rapidly, turning healthy fields into losses within weeks.

For millions of smallholder farmers, the difference between a good season and a failed harvest often comes down to timing: knowing when the risk is coming and acting early enough. Ethiopia has been at the forefront of tackling this challenge. Through the Wheat Rust Early Warning and Advisory System (EWAS), the country has built a powerful operational system combining field surveillance, weather forecasts, and disease modelling to anticipate outbreaks and guide responses. In 2021, this system helped avert a potentially catastrophic epidemic, turning a high-risk season into a successful harvest. Yet one critical piece of information remained difficult to capture: where exactly wheat is growing, and at what stage of development.

This is where the SEN4RUST project made a difference. Rather than building a new system, SEN4RUST worked alongside EWAS to test how satellite Earth observation, particularly data from the Sentinel-2 mission, could improve the system’s performance. By providing timely, high-resolution maps of wheat location and growing stage, the project assessed the added-value of a new layer of intelligence to the existing forecasting model.

The results showed clear added value. With better information on where crops are present and when they are most vulnerable, EWAS could generate more precise and actionable warnings. In practice, this means farmers and authorities could act earlier and more effectively—reducing losses and protecting livelihoods.

SEN4RUST demonstrated how Earth observation can strengthen existing national systems, enhancing their accuracy and impact without replacing them. It provides a model for how EO can be integrated into operational decision-making in agriculture, contributing directly to food security and resilience.

Wheat is a strategic crop for Ethiopia, both for national food security and for the livelihoods of millions of smallholder farmers. Over recent years, the country has made major efforts to increase production and move toward self-sufficiency. Today, Ethiopia produces more wheat than any other country in Sub-Saharan Africa, yet it remains highly exposed to wheat rust epidemics. Wheat rust disease can spread across regions and borders, carried by wind and driven by weather conditions. A single outbreak can affect hundreds of thousands of hectares. The 2010 epidemic was a stark reminder. Around 600,000 hectares were affected, and losses reached hundreds of millions of dollars. For many farmers, it meant lost income and increased vulnerability.

To manage this risk, Ethiopia developed the Wheat Rust Early Warning and Advisory System (EWAS), bringing together field surveillance, meteorology, and disease modelling. The system marked a major step forward—but it still faced a key limitation: incomplete and delayed information about crop distribution and development. In a country with millions of small, fragmented fields and multiple growing seasons, this information is difficult to obtain through ground surveys alone. Yet it is essential: wheat rust only affects crops when they are green and in a susceptible growth stage. Without this knowledge, forecasts inevitably carry uncertainty. SEN4RUST was designed to address this specific need and close this gap, by bringing satellite-derived crop intelligence into EWAS.

SEN4RUST focused on one central question: can satellite data improve knowledge of where wheat is grown and how it develops over time—and does this improve disease forecasting?

To answer this, Sentinel-2 multispectral time series were used to generate high-resolution maps of wheat distribution across Ethiopia. With frequent revisit times and spatial resolution down to 10 meters, Sentinel-2 enables identifying where wheat is grown, even at the scale of smallholder fields. In addition, time-series analysis of Sentinel-2 data enabled the extraction of crop phenology, still at 10-meters, tracking the progression of wheat through its growth stages on a weekly basis during the season. These EO-derived products provide two critical inputs:

These datasets were then integrated into the EWAS modelling framework. Previously, the system relied on more generalized or static assumptions about crop distribution. By introducing dynamic, observation-based information, SEN4RUST allowed the model to better align disease risk with actual crop conditions. The impact is a more precise forecasting capability. Risk maps become more targeted, and warnings can be issued with greater confidence. Importantly, these improvements are delivered within the existing EWAS workflow. The EO products are not standalone outputs, they are embedded into the operational system. This approach ensures that EO data is not just available, but usable and impactful.

The SEN4RUST contribution builds on the strong foundation of EWAS by adding a new observational layer from space.

The key innovation is not the individual data sources, but their integration. By linking satellite observations with ground data and models, SEN4RUST enhances the overall system performance, reducing uncertainty and improving reliability.

The impacts of SEN4RUST are substantial, particularly in terms of food security, economic resilience, and farmer livelihoods. The benefits are best understood through its contribution to decision-making.

For farmers, improved forecasts mean better timing. With more accurate information on when and where wheat is at risk, they can apply fungicides more effectively, protecting crops while avoiding unnecessary costs. Even small improvements in timing can have a large impact on outcomes. For national authorities, EO integration enables more targeted responses. Instead of broad, precautionary measures, interventions could be focused on areas where risk is highest. This will improve efficiency and ensure that resources are used where they are most needed.

The added value of EO is particularly evident in high-risk seasons. When environmental conditions favor disease spread, having accurate and up-to-date crop information becomes critical. SEN4RUST demonstrated that EWAS could benefit from EO data to anticipate and respond to such situations.

More broadly, the project shows how Earth observation can enhance existing operational systems. Rather than requiring entirely new infrastructures, EO can be integrated into established workflows—adding value in a practical and scalable way. This has important implications beyond Ethiopia. The approach could be extended to other countries facing issues with wheat rust (e.g. Nepal and Bangladesh working already with EWAS and other countries in Africa and Asia currently developing their own early warning systems). Beyond wheat rust, the methodology could probably be adapted to other crop diseases, offering a powerful tool for global agricultural resilience.