Rice ADM1 Climate–Yield Framework: Phenology-Aligned, Rice-Area-Weighted Thermal Exposure and Multi-Country Validation

This dataset supports a multi-country framework for constructing phenology-aligned, rice-area-weighted thermal exposure and evaluating its association with rice yield at the first-level administrative-unit (ADM1) scale. The archive contains derived climate-exposure datasets, statistical outputs, quality-assurance and quality-control files, and publication figures for ten rice-producing countries. Core long-panel inference focuses on China, Japan, India, Brazil, and the Philippines. Thermal exposure was derived from daily maximum and minimum air temperature and represented by mean temperature (Tmean) and diurnal temperature range (DTR). Climate exposure was spatially weighted by rice production area and temporally aligned to flowering using 15-day windows centered at −14, −7, 0, +7, +14, and +21 days relative to flowering. The archive includes country-year and ADM1/cluster-level exposure products, flowering-stage climate-trend estimates with HAC(3) inference, country-specific DTR–yield associations, cross-country Wald coefficient-equality tests, BH-FDR multiple-testing diagnostics, Brazil timing reconstruction files, harmonization audits, and publication-ready figures. The principal analytical finding is that a standardized climate-exposure construction can be transferred across rice-production systems, whereas the magnitude and direction of DTR–yield associations remain context dependent. The dataset therefore supports the distinction between a transferable exposure framework and a universal climate–yield response coefficient. Upstream meteorological, rice-area, phenological, and agricultural statistics remain subject to the terms of their respective data providers. This archive primarily distributes derived analytical products and QA/QC outputs generated for the associated study.