Remote Sensing Time Series for Modeling Invasive Species Distribution: A Case Study of Tamarix spp. in the US and Mexico

Abstract

Detecting invasive species and predicting their potential distribution are crucial to coordinate management responses. Remote sensing data are now available in several spatial and temporal resolutions and can supply environmental models with additional information. This study uses the Maximum Entropy algorithm to model the current distribution of the saltcedar (Tamarix spp.) in the US and Mexico and to identify suitable habitats, both already inhabited and not yet occupied. Tamarisk is restricted to specific habitats such as riparian zones, wetlands and agricultural or disturbed areas, which are typically not only characterized by climate. To describe vegetation phenology and thermal seasonality in these habitats, the study uses annual metrics of remotely sensed time series from 2001 to 2008 (Terra-MODIS Enhanced Vegetation Index and Land Surface Temperature) together with WorldClim bioclimatic data. By using occurrence records primarily from the US we were able to model predictive maps of tamarisk distribution correlating very well to the known distribution in the US. For Mexico, where only very few occurrence records exist, we identified potential tamarisk habitats for substantial areas in Baja California, in the states of Sonora and Sinaloa and in the Central Mexican Plateau. These predictive model results can be used to support the early detection and prevention of Tamarix spp. invasion.