Estimating colonization and invasion risk maps for Linepithema humile, in Japan

Our goal was to create colonization and invasion risk maps for the Argentine ant, Linepithema humile, based on occurrence data in Japan, by combining colonization- and invasion-related variables and spatial filters that alleviate spatial autocorrelation. With these data, we will be better able to implement surveillance and control programs. Species distribution models were generated, using the maximum entropy approach, from presence-only data collected from 12 locations. Colonization-related variables (e.g., temperature, precipitation) and invasion-related variables (e.g., urban area, distance from ports) were used as environmental variables and spatial filters that alleviate spatial autocorrelation were included at the same time. The high invasion risk area was restricted to coastal areas, whereas high colonization risk applied to a broader area. Elevation, minimum temperature, and flow accumulation were the most effective variables for predicting colonization risk, whereas urban area, elevation, and the port distance index were the most effective variables for predicting invasion risk. The invasion risk map had a higher level of accuracy than the colonization risk map. We identified those areas with a high risk of invasion in the early stages and strong propagule pressure with a model using both invasion-related variables and colonization-related variables to accurately estimate the initial invasion distributions. We found that high colonization risk areas were concentrated in the Okinawa and Ogasawara Islands; ecosystems with highly endemic ant species that are likely to have a high sensitivity to L. humile introduction. Our data will aid in strengthening both domestic and international quarantine systems to prevent such introductions. (C) 2015 Korean Society of Applied Entomology, Taiwan Entomological Society and Malaysian Plant Protection Society. Published by Elsevier B.V. All rights reserved.