Global dynamic precipitation isoscapes over three-quarters of a century

Abstract. Stable precipitation isotopes are widely used as tracers in water cycling and material transport. However, observational networks (e.g., the Global Network of Isotopes in Precipitation, or GNIP) are sparse in space and time, constraining analyses and applications in data-poor regions. To bridge this gap, static isoscapes, which interpolate monthly observations to long-term annual or climatological monthly means, have been developed and applied across disciplines. In recent years, growing interest in event-scale processes and isotope-enabled hydrological modeling has increased the need for higher-frequency information.

In this study, a global precipitation isoscape using an offline isotope circulation model (ICM) forced by JRA-3Q on a 1.25° grid covering −80° to 80° of latitude is produced. The deliverables include daily values, precipitation-weighted monthly values, and climatological monthly means for 1948–2023; the daily and monthly time series cover September 1947 to March 2024. GNIP-based validation demonstrates high skill for climatology. When all the station–month pairs are analyzed in a single regression, R² equals 0.86 for δ¹⁸O and 0.87 for δ²H, with root-mean-square errors (RMSEs) of 2.02 ‰ and 16.0 ‰, respectively. At the monthly scale, unweighted averaging of station-level metrics yields R² values of 0.47 and 0.48, with RMSEs of 2.62 ‰ and 19.9 ‰, respectively. Daily performance was quantified from same-day regional means in East Asia and from 21 GNIP stations not used in the correction; R² is typically between 0.3 and 0.6, and the RMSE is between 2 and 4 ‰ for δ¹⁸O. By comparison, d-excess exhibits lower global skill, likely reflecting sensitivity to uncertainties in the forcing and to model simplifications.

In contrast to observation-interpolated static isoscapes, this reanalysis-driven dataset delivers seamless coverage, including data-sparse regions, with explicitly characterized error properties. The data are distributed as NetCDF/CSV (daily, monthly) and GeoTIFF (climatology) and support applications from isotope-enabled hydrology and source attribution to climate-impact assessments and global water-resource analyses.