Modelling potential current distribution and future dispersal of an invasive species Calliandra calothyrsus in Bali Island, Indonesia

Abstract. Yudaputra A. 2020. Modelling potential current distribution and future dispersal of an invasive species Calliandra calothyrsus in Bali Island, Indonesia. Biodiversitas 21: 674-682.  Calliandra calothyrsus Meisn. is relatively well-adapted in abandoned areas, degraded lands, and poor nutrient soils. It tends to reproduce rapidly and be invasive in certain landscapes as it often dominates the vegetation. This study aimed to understand the potential current distribution and the population dispersal of C. calothyrsus across Bali Island using Random Forest (RF) and Maximum Entropy (MaxEnt) models. Thirteen environmental variables, including several climatic variables, topography, soil characteristics were used as predictors. The occurrence records of C. calothyrsus were obtained from direct field survey in which square plots 10 x 10 m were used to collect the population structure data. The Rangeshifter software was used to understand the population dynamic and dispersal pattern. The results showed that the two models (RF and MaxEnt) have the AUC>0.9 which means those models are excellent in predicting the potential current distribution of C. calothyrsus. Furthermore, the RF model has the TSS and Kappa value of >0.90 which means it has almost perfect agreement between the prediction and the real observation. On the other hand, the TSS and Kappa value of MaxEnt were >0.70 indicating it has a substantial agreement. The population structure in the field showed that the number of juvenile individuals dominated all plots compared to seedlings and mature individuals. The simulation analysis showed that the population tends to have bigger population in the next 50 years by dispersing throughout neighbor cells or areas in which the origin occurrence points were recorded.