Protected areas are essential for both conservation of biodiversity and provisioning of dynamic ecosystem services. In recent years, invasive plant species have had devastating impacts on the sustenance of natural habitat for the in-situ conservation of threatened species. In the present study, we investigated the distribution pattern of Chromolaena odorata and Mikania micrantha and factors governing their spread in the grassland of Manas National Park. We carried out extensive field sampling and modeled the distribution of invasives using a suite of algorithms. Model predictions differed with respect to AUC, sensitivity, specificity and TSS range, largely in classification performance. Final risk maps were produced averaging the model outputs of Random Forest and MaxEnt for both the species. Subsequently, we also emphasized that the niche/SDM modelling studies should instigate by testing a suite of algorithms for predictive ability under the specific conditions of the study since no single optimization approach would be best under all circumstances. Proximity to the road and human settlement, elevation, fire occurrence and precipitation of driest quarter were the key predictors for both the invasive.

The potential area invaded by the C. odorata was 74.87 sq. km and M. micarantha 81.82 sq. km, and a high degree of overlap found between the distributions of both the species. Invasion risk maps can be used as an early detection tool for the management of invasive species, which could help in minimizing the ecological significance and economic cost of invasions. The findings of the present study will help the forest authorities to manage, take the necessary steps in the identified area to chalk out a comprehensive strategy for control of invasive.