Predicting the distribution and abundance of invasive plant species in a sub-tropical woodland-grassland ecosystem in northeastern India

Abstract

Invasive plant species have become increasingly problematic in tropical and sub-tropical ecosystems, with the potential to decrease native plant diversity, increase fire occurrence, and cause ecosystem degradation. Numerous factors including disturbance due to fire, grazing, roads, human activities, reduction of native diversity, and soil fertility are known to influence invasibility of a habitat and/or promote the spread of invasive species. We studied invasive species distribution and abundance in a 519 km2 wildlife reserve that has sub-tropical woodland and grasslands. We sampled 134 plots of size 30 × 30 m2 and found that Mikania micrantha (a climber) and Chromolaena odorata (a shrub) were the most prominent invasive plants. We then tested the influence of eleven environmental variables that are either direct measures or proxies of resource availability, vegetation density, disturbance, and moisture stress. Using these predictors, we performed a decision-tree-based regression and prediction to test the influence of these variables on invasive species abundance and to generate distribution maps. The model had significant predictive power in the case of Mikania (R2 = 0.469) but was poor for Chromolaena (R2 = 0.056). Annual precipitation, soil phosphorus, and vegetation attributes had a significant influence in Mikania, and fire frequency had the strongest influence on Chromolaena. We could not quantify direct disturbance such as cattle grazing and resource extraction, which could add to the predictive power for these species. Given that invasive species continue to expand in range and abundance, more directed ecological monitoring and analyses are needed to manage ecosystems under the threat of invasions.