Forecasting Potential Distributions of Nonindigenous Species with a Genetic Algorithm

Publisher Website
Google Scholar
Abstract
Forecasting species' potential distributions is a leading problem for research in invasion ecology and global change biology. Given the number of freshwater organisms, especially fishes, that have been introduced intentionally and unintentionally, an urgent need exists for the development of tools for forecasting their spread. New computational techniques partially overcome the defects of earlier environment‐matching methods, including sensitivity to sparse and unbalanced data, inflated statistical significance, and asymmetrical (omission versus commission) errors. Here we apply one such method, the genetic algorithm for rule‐set production (GARP), to predict spatial occurrences of suitable environments for two nonindigenous fishes that so far have limited distribution in North America, Eurasian ruffe (Gymnocephalus cernuus) and rainbow smelt (Osmerus mordax). For these two species, we conclude that although spread is not inevitable, their eventual range in North America will be considerable if management actions fail to prevent additional introductions.

This publication has 0 references indexed in Scilit: