Model-based forecasting of bacterial black node of barley using a hierarchical Bayesian model

摘要

Bacterial black node (BBN) due to Pseudomonas syringae pv. syringae (PSS) is the most serious bacterial disease of barley in Japan. To help growers determine when to apply control measures against BBN, we developed a disease-forecasting model using a hierarchical Bayesian model (HBM) based on 29 years of data from Kagawa Prefecture (1992-2020), 19 years from Okayama Prefecture (2002-2020), and 8 from Yamaguchi Prefecture (2013-2020). The model included the number of fields with BBN in May of the previous season and the number of days at a minimum temperature (<= - 4 degrees C) in January of the current season as predictors. The model was validated using a fivefold cross-validation (CV) procedure and achieved an average accuracy of 0.713, suggesting that this model can be used to predict the BBN incidence in May of the current season. This is the first report on developing a disease-forecasting model for BBN incidence using HBM based on a total of 56 years of historical data from three prefectures.