Every Raindrop Counted: 63 Years of Precipitation Across China

China's relationship with water is one of civilization's oldest stories. The same monsoon system that feeds the rice paddies of Guangdong can unleash catastrophic floods along the Yangtze. The same orographic lift that wrings moisture from Himalayan air currents leaves the Gobi bone-dry. Understanding precipitation at fine spatial and temporal scales is not an academic exercise in China — it is an infrastructure imperative. CHM_PRE V2 provides daily gridded precipitation at high resolution from 1961 to 2024, capturing the full spectrum of China's rainfall diversity with unprecedented accuracy.

What sets CHM_PRE apart from competing products is its treatment of spatial autocorrelation — the statistical tendency for nearby locations to experience similar rainfall. Most gridded precipitation datasets interpolate station data without accounting for this spatial structure, introducing systematic biases in regions where stations are sparse. CHM_PRE V2 explicitly models these spatial dependencies, then integrates satellite estimates and reanalysis products as covariates. The result is a precipitation surface that performs measurably better in cross-validation against withheld stations, particularly in mountainous terrain where conventional interpolation struggles most.

The V2 release addresses a known limitation of the original dataset: precipitation estimation along the southern Himalayan foothills, where extreme orographic rainfall gradients can produce tenfold differences in annual totals over distances of just 50 kilometers. The adjusted estimates bring CHM_PRE into closer agreement with high-altitude gauge networks in Nepal and Bhutan, making the dataset more reliable for transboundary hydrological studies. For the 400 million people who live in China's major flood-prone river basins, the precision improvements in V2 translate directly into better early warning systems and more accurate reservoir management models.