In-season performance of European Union wheat forecasts during extreme impacts

Area of Science:

• Agricultural Science
• Climate Science
• Data Science

Background:

• Accurate in-season forecasting of European wheat (Triticum spp.) yields is crucial for food security and market stability.
• Previous assessments have not fully captured the seasonal development and accuracy of these forecasts across varying yield conditions.

Purpose of the Study:

• To evaluate the quality and in-season development of European wheat yield forecasts.
• To identify forecast biases in low, medium, and high-yielding years.
• To understand the impact of weather extremes on forecast accuracy.

Main Methods:

• Analysis of 440 wheat yield forecasts for 75 years (1993-2013) across 25 European Union Member States.
• Comparison of forecast yields with actual yields, categorizing by yield level (low, medium, high).
• Correlation of forecast errors with identified weather drivers (drought, heat, wet conditions).

Main Results:

• By July, median yield years were forecast with <2% error.
• Low-yield years were overestimated by ~10%; high-yield years were underestimated by ~8%.
• Drought or heat drivers were present in 80% of low-yield years; extreme weather events caused significant underestimations in forecast accuracy.

Conclusions:

• Wheat yield forecasts exhibit systematic biases, particularly underestimating impacts of extreme weather.
• Improved operational forecasting requires integrating near-real-time information, advanced crop modeling, enhanced Earth observations, and faster computation.
• Adapting to unprecedented climate impacts necessitates more robust and accurate yield prediction systems.