Modelling presence versus abundance for invasive species risk assessment

• Published in: Diversity & distributions 2021-12, Vol.27 (12), p.2454-2464
• Main Authors: Jarnevich, Catherine S., Sofaer, Helen R., Engelstad, Peder
• Format: Article
• Language: English
• Subjects: Abundance; Agricultural production; Biodiversity; Decision making; Geographical distribution; Grasses; Introduced species; invader abundance; invasion risk; Invasive plants; Invasive species; management prioritization; Nonnative species; RESEARCH ARTICLE; Risk assessment; species distribution model
• ISSN: 1366-9516; 1472-4642
• DOI: 10.1111/ddi.13414

Aim Invasive species prevention and management can be guided by comparisons of invasion risk across space and among species. Species distribution models are widely used to assess invasion risk and typically estimate suitability for species presence. However, suitability for presence may not capture patterns of abundance and impact. We asked how models estimating suitability for presence versus suitability for abundance aligned in their implications for risk assessment. Location Western United States. Methods We developed ensembles of species distribution models for presence and for abundance for four invasive plants. We visualized the distribution of presence and abundance in environmental and geographic space and compared model outputs using criteria relevant for decision‐making: a comparison of risk across management units for each species, and a ranking of risk among species for each management unit. Results We found good overall agreement between models of presence versus abundance in the relative risk across management units and among species. However, the area predicted to be suitable for invasive species presence was often substantially higher than the area predicted to be suitable for abundance, especially within uninvaded management units. Main conclusions Models of suitability for invasive species presence and abundance yielded similar assessments of relative risk in comparisons across space and species. In addition, we found patterns of presence and abundance in environmental space can guide modelling decisions and model interpretation. Suitability for abundance can improve relative risk assessment when abundance locations occupy a well‐defined subset of the environmental space corresponding to presence. Where abundance locations occur throughout this environmental space, as was particularly striking for Taeniatherum caput‐medusae, suitability for presence may better reflect risk of ongoing population increases and spread. This species is at risk of becoming abundant across a substantial portion of the western United States.