Prediction of invasion risk of pine wilt disease based on GIS spatial technology and MaxEnt model in western Sichuan Province of southwestern China

  Objective  In China, the pine wood nematode (Bursaphelenchus xylophilus) takes Monochamus alternatus and M. saltuarius as propagative materials, leading to catastrophic damage to forests since they infect the forest trees. Predicting the invasion risk of pine wilt disease has critical referred values for forest protection and quality improvement, and as a result, it is related to national ecological security and carbon neutralization.

  Method  Based on the data of 24 distribution points for M. saltuarius and another 55 points for dead pine trees (hosts of M. saltuarius before eclosion) as well as 20 abiotic and biotic variables in Lixian County, Sichuan Province of southwestern China, we predicted the potentially suitable distribution areas for M. saltuarius and dead pine trees using GIS analytical tool and the MaxEnt model. Furthermore, the MaxEnt jack-knife of variable importance was applied to analyze the influence of main factors on the areas of M. saltuarius and dead pine trees, respectively. Considering that the occurrence of pine wilt disease at least requires both elements (i.e., M. saltuarius and pine trees), we evaluated the invasion risk of pine wilt disease by predicting the potential occurred regions of B. xylophilus based on weighting and integrating data of the occurrence of M. saltuarius and dead pine trees generated by MaxEnt models.

  Result  Mean area values under curves of the suitable distribution areas of M. saltuarius and dead pines were 0.993 and 0.969, respectively from the MaxEnt models, which indicated that the model predictions were ideal and can be used to forecast the potential invasion risk of pine wilt disease. Assessing the invasion risk of pine wilt disease demonstrated that the risk was the highest when mean annual temperature ranged from 7.8 ℃ to 10.1 ℃ and the precipitation in the wettest season was 348 mm to 358 mm as well as within 1.5 km from the nearest residential point. The model estimated that the high-risk area of pine wilt disease was 10 616 hm², accounting for 7.1% of the total area of coniferous forests in the county, which was distributed zonally along the roads across villages and towns.

  Conclusion  Using GIS spatial technology and MaxEnt modelling can benefit the prediction of the invasion risk of pine wilt disease for the forest zones in western Sichuan Province. However, there still exists great uncertainty in the transmission and occurrence of pine wilt disease due to economic construction and climate change. Therefore, it is necessary to strengthen the monitoring of pine wood nematodes (B. xylophilus) and their propagative materials along the residential areas and roads, promoting the preventive and controlled emergency response to ensure the ecological security of forest zones in western Sichuan Province.