Fig. 1: Basic information on the global precipitation dataset and the construction of regional-scale models.

a Global continental division into four climatic regions based on the annual average precipitation volume from the Intergovernmental Panel on Climate Change dataset for the years 1891-2018. b Daily average precipitation volume from CPC-Global data for the years 2015–2018. c Global ASTER GDEM elevation data. d Spatial distribution of ground station density as of December 31, 2018. e Distribution of global spatiotemporal precipitation data imputation and correction model count. Original data provided in Supplementary Table 3. f Correlation analysis between topographic factors and precipitation volume, tested for significance at 0.01. Original data provided in Supplementary Table 4. g Optimized Clustering of Spring Precipitation Characteristics in the Northern Hemisphere. The subplot illustrates the relationship between the number of clusters (X-axis) and both the sum of squared errors (SSE, left Y-axis, in blue) and the silhouette coefficient (SC, right Y-axis, in red). The optimal number of clusters for each region is indicated by the SSE “elbow” point and the SC peak within each corresponding subplot. Northern hemisphere seasonal divisions: March-May as Spring, June-August as Summer, September-November as Autumn, and December-February as Winter⁶⁹, showcasing regional and seasonal optimization of precipitation pattern clustering.