Predicting Invasive Plants: Prospects for a General Screening System Based on Current Regional Models
- Cite this article as:
- Daehler, C.C. & Carino, D.A. Biological Invasions (2000) 2: 93. doi:10.1023/A:1010002005024
Screening systems for predicting invasive plants have been independently developed for the non-indigenous floras of North America, the South African fynbos, and Australia. To evaluate the performance of these screening systems outside the regions for which they were developed, we tested them for the non-indigenous flora of the Hawaiian Islands. When known invasive plant species in the Hawaiian Islands were evaluated using the North American and Australian systems, 82% and 93% of the species were predicted to be invasive, respectively, and the remainder were classified as requiring further study. The South African fynbos system correctly predicted only 60% of the invasive species in the Hawaiian Islands. All three screening systems correctly classified a majority of the non-invaders as non-invasive. The Australian system has several advantages over the other systems, including the highest level of correct identification of invaders (>90%), ability to evaluate non-woody plants, and ability to evaluate a species even when the answers to some questions are unknown. Nevertheless, with the Australian system, a large fraction of species known not to be invasive were recommended for further study before importing, so there remains room for improvement in identifying non-invasive species. Based on our results for the Hawaiian Islands and a previous evaluation in New Zealand, the Australian system appears to be a promising template for building a globally applicable system for screening out invasive plant introductions.