Future Spatial Prediction of Invasive Plant Merremia peltata in Indonesia

Abstract

Merremia peltata is a woody vine that considered as invasive plant. It causes serious threats to the regeneration of native plant. Many strategies to control its invasiveness have been continuously implemented, but climate change causes complicated problems. This study aimed to understand the impact of climate change to its current and future distribution. Ensemble model by combining Random Forest and Support Vector Machine was used to predict its distribution. There were 33 occurrence records derived from Global Biodiversity Information Facility and reliable scientific journals. The 19 climatic variables were tested using a multicollinearity test. AUC and TSS were used to evaluate the model. Multicollinearity test using 0.7 threshold produces 7 selected climatic variables. AUC and TSS have the value >0.80 indicating the model has good performance. The predictive future distribution map shows that its distribution shifts to other regions in 2070 compared to the predictive current distribution. The predictive current distribution covers 30.4 % of the total land areas of Indonesia, while the predictive future distribution (RCP 4.5) covers 28.12 % and RCP 8.5 covers 23.59 %. It indicates that the suitable geographic distribution areas of M. peltata reduces around 3-7 % of the total areas of Indonesia in 2070.