A high-quality four-decadal (1981-2024) dataset of soybean phenology records from 43 agrometeorological stations across Northeast China

This study introduces a high-quality, long-term dataset (1981–2024) of soybean phenology records from 43 national agrometeorological stations across Northeast China, the nation's most critical commercial grain base. The dataset captures the timing of 8 pivotal phenological periods (sowing, emergence, three leaf, branching, flowering, pod formation, seed filling, maturity) and the durations of 6 pivotal growth lengths (sowing—three leaf, three leaf—flowering, flowering—pod formation, pod formation—seed filling, seed filling—maturity, sowing—maturity). A multi‑step quality assurance procedure was applied, integrating automated consistency validation and manual expert review to resolve ambiguous entries, thereby ensuring robust data reliability. Analytical approaches included kernel density estimation for characterizing event distributions and simple linear regression for evaluating temporal trends across decades. The published archive contains 656 linear regression images (JPEG) and two core data tables (XLSX), amounting to 648 MB with a compressed size of 398 MB. Organized hierarchically by province and station, this collection offers a valuable empirical foundation for investigating climate‑related alterations in crop development, refining the calibration and evaluation of mechanistic crop models, and supporting the design of improved agronomic management and regional climate adaptation plans.