Presence–absence and environmental predictor data for modelling Triadica sebifera invasion risk on Miyajima Island, Japan

This dataset contains field-based presence–absence point data and environmental predictor values used to model the current invasion risk of the invasive alien tree Triadica sebifera on Miyajima Island, Hiroshima Prefecture, Japan. The dataset was prepared for the manuscript entitled “Disturbance and anthropogenic edges drive invasion risk of Triadica sebifera in a protected island ecosystem: evidence from Landsat-based random forest modeling”. The deposited files include presence–absence point data, predictor values at survey points, and spatial block assignments used for random forest modelling and spatial cross-validation. Environmental predictors include Landsat-derived vegetation indices, topographic variables, disturbance history, and anthropogenic variables. Publicly available source datasets used to derive environmental predictors include Landsat 8/9 surface reflectance data, NASADEM, and Urban Land Use Fine Mesh Data from the National Land Numerical Information of Japan.