Long-Term and Short-Term Forecasting of Oriental Fruit Moth (Grapholita molesta) Trap Catches from Apple Orchards in South Korea Using Time Series Models

, is a major agricultural pest causing significant economic loss of apple growers in South Korea. This study demonstrates the application of time series models for describing the national and regional patterns of OFM occurrences in the last decade and for forecasting future OFM occurrences. The seasonal autoregression integrated moving average (SARIMA), Prophet, and vector autoregressive (VAR) models are compared for both long- and short-term predictions. The analysis shows that short-term predictions are more reliable than long-term predictions for the number of OMF trap catches, and the multivariate time series model does not necessarily provide better predictive performance with province-level aggregated data. Though the Prophet and VAR model fits bimonthly province-level data better than the SARIMA model, the VAR model shows poor predictive performance, and the SARIMA model showed as or more reliable predictions than the Prophet model in this study. This study presents both the potential and challenges for establishing a Smart Integrated Pest Management (IPM) system capable of monitoring and predicting OFM occurrences and implementing regional pest control strategies. The usefulness of time series analysis can be leveraged by frequent orchard-level data reporting, pest management records, and precise local environment information.