Abstract
This study evaluates climatic and anthropogenic drivers influencing habitat suitability of invasive Cuban knight anole (Anolis equestris) in Florida and assesses their potential impact on three species of threatened invertebrates due to habitat overlap. We developed species distribution models (SDMs) using eight algorithms to evaluate habitat suitability across the native and invasive range of A. equestris. We generated ten independent pseudo-absence sets at a 1:1 ratio with presences and implemented a 10-fold cross-validation scheme. Predictor variables included effort, climatic, topographic, urbanization, and vegetation indices. We trained algorithms on 70% of the data, validated on 30%, constructed both algorithm-specific and global ensembles. The best-performing model was used to assess variable importance and predict habitat suitability across regions. Random Forest (RF) demonstrated the best overall performance (Florida: BI = 0.98, TSS = 0.91; Cuba: BI = 0.89, TSS = 0.74) and was used for subsequent analyses. When projected against an independent dataset with standardized effort, the model retained discriminatory power (TSS = 0.53; BI = 0.59), indicating generalizability. Mean diurnal range was the most influential predictor overall, while urbanization (e.g., settlement model grid) was more important in Florida. NDVI and precipitation of the driest month had greater influence in Cuba. Predicted habitat suitability at occurrence locations of the endangered Florida tiger beetle (Cicindelidia floridana, x̄ = 0.86), Florida tree snail (Liguus fasciatus, x̄ = 0.58), and endangered Schaus' swallowtail butterfly (Papilio aristodemus, x̄ = 0.53), suggest potential overlap. These findings emphasize the role of urbanized habitats in facilitating invasion and provide a data-driven framework for conservation management and mitigation.