Structured Evidence Data and Reproducibility Materials for Machine Learning in Climate-Change Hydrology

This dataset accompanies a structured critical review of machine learning in climate-change hydrology. It contains a path-sanitized 146-record evidence matrix, proposition-to-case traceability, deterministic consistency and limited-text sensitivity checks, purpose-based selection and climate-use coding rules, analytic exclusions, figure source data, and a standard-library validation script. The corpus was selected for conceptual coverage rather than field-prevalence estimation. No publisher PDFs or copyrighted article full text are redistributed. Code in scripts/validate_release.py is additionally licensed under the MIT License. Funding was provided by the Zhejiang Provincial Department of Education (Y202558418) and Zhejiang Shuren University (2023R007).