Abstract

Predicting suitable habitat of an invasive species can aid resource managers in time-efficient and cost-effective monitoring and removal. Using the Software for Assisted Habitat Modeling, we tested four model algorithms to forecast potential range expansion of a new invader, Neopomacentrus cyanomos (regal demoiselle), in the Greater Caribbean. Models were built using independent training and testing data that contained presence and absence points observed in both the native and invasive ranges and a suite of environmental predictors. We compared models developed using native occurrences, non-native occurrences, and a combination of both. The Generalized Linear Model algorithm for all models performed similarly and ranked the same top three most important predictors to determine the distribution of N. cyanomos. Our model results suggest that suitable habitat for this species is relatively shallow (<50 m), includes both brackish and normal ocean salinities, and has low to moderate current velocity. The results also suggest that N. cyanomos could continue to expand its distribution into most tropical and sub-tropical coastal regions in the western Atlantic Ocean. Due to the lack of extensive natural reefs, the current invasion front in the Florida Panhandle may be an effective area to slow the spread of this species to South Florida through removal of individuals on artificial reefs. If the spread continues, South Florida may act as a gateway to islands in the northern Caribbean, as seen for lionfish. These findings are informative for raising awareness of this new invader and identifying Caribbean islands at risk for future range expansion.

PDF Preview