Development of Two Early Forecasting Models for Predicting Incidence of Rice Panicle Blast in China

Area of Science:

• Agricultural Science
• Plant Pathology
• Computational Biology

Background:

• Rice blast, caused by the fungus Magnaporthe oryzae, is a devastating disease affecting rice production worldwide.
• Early forecasting of rice panicle blast (PBx) is essential for effective disease management and minimizing yield losses.

Purpose of the Study:

• To develop accurate early forecasting models for rice panicle blast (PBx).
• To identify key factors influencing PBx incidence, including historical data, environmental conditions, and crop management practices.

Main Methods:

• Analysis of relationships between PBx and preceding crop PBx, susceptible variety acreage (SVC), altitude, and weather data.
• Development of a logistic model to predict the presence (PBx > 0) or absence (PBx = 0) of rice panicle blast.
• Development of a two-step hurdle model combining logistic and regression approaches for predicting PBx severity.

Main Results:

• Preceding crop PBx, SVC, altitude, and specific weather conditions (soil temperature, rainfall) 120-180 days before PBx date were significant predictors.
• The logistic model achieved 78.39% accuracy in predicting PBx absence/presence.
• The two-step hurdle model demonstrated a low prediction error (<6.16% for 90% of records).

Conclusions:

• The developed forecasting models provide a reliable tool for early prediction of rice panicle blast.
• These models can support informed decision-making for rice growers and policymakers in disease management strategies.
• The study offers a foundation for further research into rice blast epidemiology and forecasting.