2026-10-10T12:14:16Z https://www.scidb.cn/oai

10.57760/sciencedb.32629 2026-07-13T16:16:40Z

2026-07-13 High-resolution (0.125°) monthly predictions of precipitation δ²H and δ¹⁸O over China mainland for 1990-2020 doi:10.57760/sciencedb.32629 en High-resolution (0.125°) monthly predictions of precipitation δ²H and δ¹⁸O over China mainland for 1990-2020, together with 30-year climatologies. The dataset supports research in hydrology, ecology, paleoclimate, archaeology, isotope geochemistry and related fields.Data files:data/H2_value_1990‑2020.nc Monthly δ²H time = 372, lat = 425, lon = 521data/H2_value_climatology_1990‑2020.nc 30‑year mean δ²H lat = 425, lon = 521data/O18_value_1990‑2020.nc Monthly δ¹⁸O time = 372, lat = 425, lon = 521data/O18_value_climatology_1990‑2020.nc 30‑year mean δ¹⁸O lat = 425, lon = 521 *Approximate on‑disk sizes with NetCDF‑4 compression enabled.Common metadata Spatial extent : 2° N – 55° N, 72° E – 137° E Grid : 0.125° regular (EPSG:4326) Land mask : Applied (oceans → NaN) Created on : 2025‑06‑07 15:30 UTCLicence & citation Data The four NetCDF files in "data/" are released under the Creative Commons Attribution 4.0 International (CC BY 4.0) licence. You are free to share and adapt the data for any purpose, provided you give appropriate credit.Disclaimer: The generated datasets and source code are provided "as is" strictly for academic research and broad-scale reference purposes. It should be emphasized that the core scientific contribution of this study lies in the interpretable machine-learning methodology, and the reconstructed gridded datasets serve primarily as a byproduct demonstrating the capability of this framework. Because the datasets are derived from statistical models trained on sparse and temporally intermittent observations, inherent uncertainties remain. This is particularly true for topographically complex or observation-scarce regions (e.g., the Tibetan Plateau), under-sampled winter months, and extreme fractionation conditions. Users are advised to interpret single-month or pixel-level local values with caution. We strongly recommend that prospective users thoroughly read the original manuscript to fully understand the methodological assumptions, validation contexts, and limitations before deciding whether to use the datasets. The authors assume no liability for any specific conclusions, subsequent applications, or downstream scientific disputes arising from the secondary use or independent interpretation of these materials. d2H; d18O; CHINA; Piso_AI WANG TIAN RESTRICTED dataset http://www.doi.org/10.1016/j.accre.2026.05.014 Science Data Bank