Risk monitoring of pine wilt disease based on semi-dynamic spatial prediction in South Korea

Pine wilt disease (PWD), caused by Bursaphelenchus xylophilus, is the deadliest disease affecting pine trees, and causes severe economic and ecological damage in South Korea. Therefore, monitoring PWD is a national campaign necessary for timely control of the disease. We aimed to develop a model that predicts PWD on a monthly and to use this model to build information that can be utilized for practical monitoring. This study developed a semi-dynamic species distribution model to predict the monthly probability of PWD occurrence in South Korea, incorporating climate and anthropogenic factors along with monthly PWD occurrence data. The model was further refined by classifying risk levels across administrative districts, making it applicable for practical monitoring. Additionally, an ensemble model was created by integrating monthly PWD predictions with host distribution data. This approach identifies the most vulnerable areas at risk of PWD outbreaks, offering a targeted strategy for disease management and prevention. The results showed the highest likelihood of PWD occurrence around actual outbreak areas; however, monthly variations in disease occurrence areas were observed. Notably, owing to vector activity, the potential for spread to areas where outbreaks had not yet occurred was the highest during the summer season. Additionally, because factors contributing to PWD vary by season, monitoring should be conducted monthly, whereas the monitoring map identifies areas that require intensive management throughout the year. This study not only provides the foundational data necessary for establishing practical monitoring strategies for PWD but also offers an approach for the semi-dynamic prediction of species distribution modeling based on monthly data. These methods are expected to be useful in developing spatial prediction and monitoring strategies for forest pests and diseases over time, which are relatively limited in this field. [Display omitted] •Monthly distribution of pine wilt disease was predicted by semi-dynamic model.•The risk levels by administrative district were classified for systematic monitoring.•High-risk areas were identified by linking host distribution with an ensemble model.•Monthly changes in disease occurrence areas and major causes were examined.