Using individual-based models to develop invasive species risk assessments by predicting species habitat suitability.
Abstract
Preventing invasive species establishment is a global conservation priority, yet limited management resources oftentimes restrict sites to target for prevention or monitoring. Risk assessments based on habitat suitability can identify sites most vulnerable to invasion that should be prioritized for preventative actions. Since habitat suitability is the result of interactions between environmental and organismal attributes, analyses should incorporate individual variability in demographics expected in an invasive population. Individual-based models (IBMs) can predict habitat suitability by accounting for interactions between environmental conditions and individual-level variability. We developed an IBM to predict suitability of rivers in the northern United States to the invasive fishes silver carp (Hypophthalmichthys molitrix) and bighead carp (H. nobilis) and explored the projected effects of climate change on habitat suitability. All rivers supported adult survival, although complete survival of all adult demographics and positive growth only occurred in approximately 45% (17 of 38) of rivers for silver carp and 26% (10 of 38) of rivers for bighead carp. Only the largest individuals at the time of introduction survived in rivers where adult mortality occurred. Most rivers were unsuitable for young-of-year (89% and 92% of rivers for silver carp and bighead carp respectively). Climate change simulations had relatively little effect on adult habitat suitability but resulted in up to four times the number of rivers being suitable for young-of-year by the late-21st century and greatly extended the viable spawning season by up to an additional 65 days for silver carp and 77 days for bighead carp. Synthesis and applications. Our approach of using individual-based models as a risk assessment tool informs proactive conservation planning by identifying sites for invasive species early detection monitoring, promoting the development of contingency response plans and allowing for proactive prevention efforts. Model predictions also provide specific management guidance regarding the size and life stages to target for monitoring efforts, which capture gears to use, and the most effective time to sample for early detection monitoring.