Field-plot dataset of Fusarium head blight (FHB) severity in wheat: fungicide-untreated small plots across the United States

This dataset contains small-plot field observations of Fusarium head blight (FHB) severity in wheat from fungicide-untreated research plots across multiple U.S. states. It was compiled from cooperators in the Integrated Management Coordinated Project (IM-CP) of the U.S. Wheat & Barley Scab Initiative (USWBSI) and is intended as a benchmark dataset for modeling and prediction of FHB risk in U.S. wheat production systems. The dataset includes plot-level FHB assessments (FHB index) along with basic trial information such as year, state, anthesis (flowering) date, and cultivar resistance level. All observations in this dataset are from fungicide-untreated plots. Observations span a range of environments and epidemic intensities, making the data suitable for classical statistical analyses, machine-learning approaches, and method comparison studies in plant disease epidemiology. Original data contributed by project cooperators were cleaned, error-checked, and then augmented with associated environmental (weather-based) variables. Environmental variables may include, for example, daily or hourly period-aggregated weather measures (e.g., mean daily temperature or maximum relative humidity in a 5-day window) relative to anthesis, which allows for the exploration of relationships between weather and FHB epidemics. This dataset has supported peer-reviewed studies on FHB risk modeling and weather–disease relationships (see “Is referenced by” in the Related works section for associated journal articles). Users can reuse the dataset to develop new predictive models, compare modeling frameworks, test sampling and study design ideas, or integrate FHB data into broader analyses of wheat disease risk, yield loss, and management strategies. A README.txt file in this record, as well as the codebooks, provide detailed documentation of the data structure, variable definitions, units, and any data-processing steps applied before release. Users are encouraged to consult those files before analysis.