This repository contains the relevant code and data for the paper Long-term observation of global nuclear power plants thermal plumes using Landsat images and deep learning (Wei et al, 2023, Remote Sensing of Environment). Specifically, the U-Net.zip file includes the associated codes for segmenting surface thermal plumes from nuclear power plants along the global coasts and the Great Lakes by using the U-Net model integrated with prior knowledge. The GCNT-Plume.zip file includes the occurrence footprints of core area plumes (the occurrence_all folder), raw water temperature increment (WST) images (the delta folder), mixed area plumes and annotations (the extractWithLocation folder), model-predicted core area plumes (the prediction*_* folders), the mixed/core area plumes and background areas in shapefile format (the sampleAnnotation* folders), and location information (the location.xlsx table). Please refer to the README.md file in the U-Net.zip file for more detailed information.