Monitoring and predicting the potential distribution of alien plant species in arid ecosystem using remotely-sensed data

Human activities cause introduction of alien species to new areas, which may cause serious problems and threats to biodiversity and ecosystem services. Omayed biosphere reserve (OBR) at the northwestern desert of Egypt, recognized for its rich flora, has recently encountered new human-induced disturbances. The current study sought to identify alien species that might have invaded OBR as a result of such activities; and to estimate the area affected by alien species using remotely-sensed data. Field surveys of 300 sampling plots representing different habitats were conducted from spring 2011 to spring 2014, whereby occurrences of alien plant species were recorded. We evaluated the use of combined environmental and remotely-sensed data for developing predictive distribution models for 21 alien species recorded in the area using Maxent modeling technique. The inclusion spectral predictors with other environmental predictors has improved the performance of the distribution models (mean test AUC 0.88 +/- 0.06) for most of the alien species. Environmental variables that contributed the most to prediction of alien species distribution included soil, elevation and slope; in addition to predictors that represented disturbance proxies such as distance from irrigation canals and distance from roads. The results revealed that more than 40% of the area is predicted to be infested by at least one alien species. Results also manifest the merit of incorporating remotely-sensed data in prediction of alien species distribution, which hold promise for development of proactive management approaches to identify and control areas of high infestation of alien species and prevent further spread of invasive plants.