Integrating global and regional niches with nested hierarchical SDMs reveals niche features and invasion-phase shifts under climate change in oceanic-island grass invaders
Year: 2026
Type: article
Abstract: On oceanic islands, biological invasions challenge species distribution models (SDMs) because estimated niches depend on spatial scale and invasion phase. Islands further complicate SDMs because their limited extent truncates the local environmental space, steep elevational gradients create strong microclimatic heterogeneity, and invasion data are often sparse and spatially biased. We present a Nested Hierarchical SDM (NHSDM) framework that explicitly links global and regional niches to improve model transferability and management relevance. We model three invasive C4 grasses, Cenchrus setaceus , Melinis repens subsp. repens , and Setaria parviflora , combining worldwide occurrences with a high-resolution regional workflow for the Canary Islands. NHSDMs comprise global and regional ensembles, plus two integrative products: (i) a covariate-augmented regional model that includes the global prediction, and (ii) a multiplicative consensus. To diagnose scale effects, we introduce two metrics: the Variable Importance Deviation Index (VIDI), which quantifies shifts in predictor importance, and the Gradient Response Deviation Index (GRDI), which summarises differences in the shapes of predictor–response curves. We project suitability under current climate and late-century scenarios (SSP3-7.0, SSP5-8.5; 2071–2100) and estimate spatial overlap among species. NHSDMs reveal species-specific scale dependence: C. setaceus shows broad warm-xeric suitability and a relatively low VIDI (consistent responses across scales), while M. repens and S. parviflora display mesic-montane affinities with higher VIDI and GRDI, indicating stronger scale-driven shifts, particularly for precipitation seasonality. Despite modest environmental niche overlaps, spatial co-occurrence is projected to increase, especially under SSP5-8.5, suggesting emerging multi-species impact hotspots where coordinated control may be most cost-effective. By translating future suitability into management-oriented risk metrics, our approach supports spatial prioritisation of surveillance and control; its main value lies in guiding decisions that reduce impacts on insular biodiversity and protected habitats.
Source: Biodiversity and Conservation
Institution: Universidad de La Laguna
Language: English
Subfield: Ecological Modeling
Field: Environmental Science
Domain: Physical Sciences
SDGs: Life on land, Climate action
Open Access status: hybrid
Estimated APC paid: $3,990
Funder: Universidad de La Laguna
OpenAlex ID: W7162685873