West Nile (WNV) is a zoonotic mosquito-borne virus with an expanding geographical range and epidemic activity worldwide. Computational studies have contributed to the understanding of factors driving WNV occurrence, particularly in North America and Europe, providing invaluable insights towards surveillance and preparedness. Archipelagos have largely been overlooked, despite the risks WNV poses to unique local avian species and human populations. In this study, we apply a machine learning-based ecological niche approach, trained on WNV occurrence and (a)biotic factors from Portugal, Spain, and Italy, to estimate ecological suitability for WNV occurrence across several Atlantic archipelagos. The approach gives weight to the temporal dimension, moving beyond conventional spatial suitability estimations, and generating novel insights on seasonality both for Europe and the archipelagos. For Portugal, Spain and Italy, modelling results align with previous findings on spatial hotspots and (a)biotic drivers of WNV occurrence, while further unraveling properties of at-risk human populations within dynamically suitable land areas. For Atlantic archipelagos, results constitute a novel and detailed perspective on local ecological suitability for WNV occurrence, providing a data-driven framework that identifies spatial hotspots, defines seasonal patterns and quantifies the local population at risk. The synthetic data generated in this study supports the development of targeted preparedness, surveillance and mitigation plans tailored to the unique ecological and seasonal dynamics of each region under study.