Potential Occurrence Risk Prediction of Sudden Oak Death Under Different Future Climate Scenarios Based on SVM Model

Sudden Oak Death (SOD), a kind of plant disease, is caused by Phytophthora ramorum. It was discovered in 1993 for the first time, has a wide range of host plants, rapid spreading, serious harmful consequences. The outbreak of SOD is influenced by various factors. Thus there are obvious spatial variations in its distribution due to the influence of different environmental factors. In this study, an integrated datasets of the outbreak points and the associated environmental variables at global and regional scales were collected and processed. The Support Vector Machine (SVM) model was adopted to predict the potential occurrence risk of SOD under future climate scenarios. In addition to the traditional bioclimatic variables, Leaf Area Index (LAI) was introduced into the potential risk prediction models to achieve the forecasting of SOD in China. Four future climate scenarios of RCP2.6, RCP4.5, RCP6.0 and RCP8.5 were considered and compared. The optimal thresholds were determined for different climate scenarios and different years. The predictive accuracies were assessed using the indices of OPS, Sensitivity, Specificity, Kappa coefficient and AUC. Areas of the potential invasion risk of SOD under different climate scenarios in China were analyzed. Results showed that, under the future climate scenarios in 2050 and 2070, Yunnan, Sichuan, Guizhou, Tibet and Chongqing all have high risks. This study could provide the long-term early warning about the outbreak and invasion risk of SOD, serving for the prevention and treatment of forest diseases as well as ensuring the forest ecological security globally and nationally.