The publication "Invaders from Islands: Thermal Matching, Potential, or Plasticity?" is comprised of data derived from multiple datasets. These datasets include climactic and temperature variables used in ecological niche models for predicting suitable habitat for Leiocephalus carinatus in its invaded and native ranges. Also included are thermal tolerance measurements assessed in March 2020 for Leiocephalus carinatus from Key Largo and Cocoa Beach, USA. Additionally, we include weatherstation data for local minimum and maximum ambient temperatures for the experimentally assessed populations at short (months) and long (decades) timescales. The dataset "curly_thermal_limits" describes thermal tolerance data for Leiocephalus carinatus captured in March 2020 from two nonnative populations in Florida. The dataset "pre_study_ambient_temps_long" contains ambient minimum and maximum temperatures derived from weather stations near Cocoa Beach and Key Largo, Florida, spanning from 1991 to 2020. The dataset "pre_study_ambient_temps_short" contains ambient minimum and maximum temperatures derived from weather stations near Cocoa Beach and Key Largo, Florida, spanning from October 1 2019 to March 8, 2020 (the last day of lizard thermal lizard sampling for this study). There are three occurrence datasets with GPS locations used to develop ecological niche models representing all points, invasive points, and native points, respectively. (ALL_occurrences.csv , INVADED_occurrences.csv, NATIVE_occurrences.csv). There are two datasets describing background points (GPS locations) used in ecological niche models for the invaded and native ranges, respectively (background_points_invad.csv , background_points_NR.csv ). There is one dataset (thresholded_differences.csv) containing thresholded relative suitability data extracted from two ecological niche models. There are three sets of six raster files used to develop ecological niche models (described in detail in Step 3), for native range (files ending with "n"), full model (invaded + native range; files ending with "f") and model projection area (files ending with "p"). In addition to data, we included scripts and model output as part of this data release. One zipped folder (R_scripts) contains various R scripts used in the analysis, and another zipped folder (Ensemble_model_output) contains output files from ensemble model analyses using BIOMOD (a free open-source package implemented in R).