Modelling Parthenium hysterophorus invasion in KwaZulu-Natal province using remotely sensed data and environmental variables

Abstract

Globally, infestations of invasive plant species impact negatively on the environment, livelihoods and local and national economies. Therefore, understanding their potential distribution is paramount. In this study, we sought to model habitats susceptible to Parthenium (Parthenium hysterophorus) invasion using environmental and Sentinel 2 multi-spectral imagery variables in a Maximum Entropy (MaxEnt) environment. Using occurrence points we tested the utility of (i) spectral bands (ii) derived vegetation indices and (iii) the combination of spectral bands and vegetation indices. Area under curve (AUC) values in MaxEnt were used to evaluate the performance of the models. Results show that the combination of environmental variables and spectral bands yielded superior AUC value of 0.976. Sentinel 2 red-edge band (705 nm) and the normalized red edge vegetation indices were the most influential variables in predicting habitat susceptible to Parthenium invasion. These results illustrate the potential of Sentinel 2 MSI and environmental variables in predicting Parthenium invasion.