Invasive plant pests have a massive social impact in Europe, costing billions of euros every year. Accurate predictions of invasive hotpots—areas with the highest risk of incurring invasions—could help focus efforts on prevention and control.
For this study, researchers developed a machine-learning method to create risk maps of pest introductions. Rather than focussing on a single species, the authors trained their model on a dataset covering 243 invasive species.
Following cleaning and thinning the final dataset feeding the authors' "elastic-net" algorithm consisted of 170,460 GBIF-mediated presence records enhanced with a rich feature set covering climate, soil, erosion risk, landcover and water indicators, as well as population and road densities, anthroprogenic pressure, distance to cities and ports, and nightlight radiance as a proxy for GDP.
The analysis revealed the highest risk of introductions in the BeNeLux states, northern Italy, the northern Balkans and the United Kingdom, with areas around container ports like Antwerp, London, Rijeka and Saint Petersburg meriting particular attention as potential invason hotspots.