Invasive Species Compendium

Detailed coverage of invasive species threatening livelihoods and the environment worldwide

Abstract

Remote sensing-derived bioclimatic variables for modeling invasive Prosopis juliflora distribution in a region of limited meteorological stations.

Abstract

Accurate modeling of invasive species in areas of limited distribution of meteorological stations is challenging. In this regard, climate records from satellites are good alternatives. However, the accuracy of these datasets needs to be validated and their performance should be evaluated. Hence, this study aimed at evaluating the performance of four satellite-derived bioclimatic variables for modeling invasive Prosopis juliflora distribution in the dryland ecosystem of Ethiopia. Accordingly, WorldClim1.4 bioclimatic variables were used as a baseline to evaluate their performance, while, gauge-derived bioclimatic variables were used to validate all tested datasets. A total of 240 species occurrence and absence points were used to train and test the performance of all models using the Random Forest algorithm. Accordingly, satellite-derived bioclimatic variables provide better performance with an area under the curve of 0.64 and true skill statistics of 0.4 compared to a reanalysis and WorldClim1.4 derived bioclimatic variables. However, the lower performance of satellite datasets indicates the importance of its integration with other variables.