Prompt-defined AI-agent workflow and derived data for climate-sensitivity assessment of the Remote Sensing Ecological Index in Qinghai Province, China (2000–2024)

This dataset accompanies the study “A human-supervised AI agent workflow for detecting climate sensitivity of the Remote Sensing Ecological Index in high-elevation landscapes.” It provides the complete computational reproducibility record for the workflow, including numbered Codex prompts, executed Google Earth Engine and ArcPy/Python scripts, and Codex-generated runner files and processing logs. The materials document the full sequence used to construct annual RSEI, prepare climatic variables, conduct trend analysis, GeoDetector attribution, pixel-wise climate-sensitivity modelling and EcoBound threshold detection for Qinghai Province during 2000–2024. Original MODIS and climate products are publicly available from their respective providers and are not redistributed. Users can reproduce or adapt the workflow by updating local project paths, study-area inputs, target years and the documented Google Earth Engine asset references.