We developed a 0.01-degree gridded precipitation dataset of Japan based on historical observation datasets covering 1926 to 2020. Historical observations conducted by the Japan Meteorological Agency and other Japanese bureaucratic agencies were spatially interpolated using the inverse distance weighting method at daily and hourly temporal resolutions. Optimal parameterization for our interpolation process was selected by comparing interpolated results of various parameter combinations with precipitation observation conducted by the University of Tokyo Forests. We conducted cross-validation for over 1,000 stations with sufficient data throughout our data period and verified our product can reproduce the temporal variability of local precipitation. The strong points of our precipitation dataset are its high spatiotemporal resolution and the abundance of point precipitation source data. We expect our dataset to be highly relevant to various future studies as it can serve multiple purposes such as forcing data for hydrological models or a database for analyzing the characteristics of historical rainfall events.
© 2022. The Author(s).