Dataset for "Zero-Shot Precipitation Downscaling Across Global Climate Regimes Using a Foundation-Style Deep Learning Framework"

This dataset contains all processed data, analysis scripts, trained model files, and post-processing utilities used in the study “Zero-Shot Transferability of Deep Learning Precipitation Downscaling Across Global Domains and Century-Scale Climate Projections”. The materials included here support the full reproducibility of the results presented in the main article and Supplementary Information.本数据集涵盖了题为《跨全球区域与百年尺度气候预估的深度学习降水降尺度零样本（Zero-shot）可迁移性》的研究中所使用的全部预处理数据、分析脚本、训练好的模型文件及后处理工具。本数据集收录的材料可完全复现正文中与补充信息里呈现的所有研究结果。