NorthChinaWheat30: A 30 m Odd-Year Winter Wheat Planting Distribution Dataset in the North China Plain from 2005 to 2025

The WDI was integrated with the Time-Weighted Dynamic Time Warping (TWDTW) similarity measure and PCA-reduced NDVI time-series features to construct the TPW feature set. These features were then input into a radial basis function support vector machine to produce winter wheat planting distribution maps for the odd years from 2005 to 2025.NorthChinaWheat30 can support winter wheat area estimation, planting frequency analysis, spatiotemporal change monitoring, agricultural spatial management, food security assessment, and cultivated land protection in coal-grain overlapping areas.

Accuracy assessment based on independent samples, ablation experiments, comparison with existing winter wheat products, and statistical yearbook validation indicates that the dataset has good classification accuracy and temporal continuity. The multi-year average overall accuracy and F1 score for the odd-year mapping results are 0.926 and 0.934, respectively. The average coefficient of determination between the remotely sensed planting area and municipal-level statistical yearbook records is 0.908, with an average RMSE of 55.19 thousand hectares.