How to model a new invader? US-invaded range models outperform global or combined range models after 100 occurrences

Area of Science:

• Ecology
• Invasive Species Biology
• Biogeography

Background:

• Invasive species pose significant economic and ecological threats.
• Species distribution models (SDMs) are vital for mapping invasive species.
• Current methods often misrepresent invasive distributions due to niche shifts and data issues.

Purpose of the Study:

• To investigate the optimal data source for SDMs of invasive species.
• To identify a "tipping point" where invaded range data outperforms global data.
• To improve predictions of potential invasive species distributions.

Main Methods:

• Utilized time series records of 13 invasive plant species in North America.
• Compared SDMs built with invaded range, global range, and combined occurrence data.
• Analyzed model performance across different invasion stages.

Main Results:

• Early invasion stages: models using both invaded and global data performed best.
• After ~100 observations: US-invaded range models significantly outperformed others.
• Higher data quality at the invaded-range scale contributed to improved model performance.

Conclusions:

• For newly invasive species, modeling potential distribution using only invaded range data becomes more accurate after limited observations.
• Disregarding global data improves accuracy once sufficient invaded-range data is available.
• This study provides a robust approach for modeling novel invasive species distributions.