Li et al. (2026) An observation-based daily gridded precipitation dataset for forecast verification

Identification

• Journal: Scientific Data
• Year: 2026
• Date: 2026-09-22
• Authors: Qiang Li, Tongtiegang Zhao
• DOI: 10.1038/s41597-026-08145-8

Research Groups

• Southern Marine Science and Engineering Guangdong Laboratory (Zhuhai), School of Civil Engineering, Sun Yat-Sen University, Guangzhou, China

Short Summary

This paper develops an observation-based daily gridded precipitation dataset for forecast verification, covering the period from 2020 to 2025. The dataset is created by integrating five global station-based observational datasets and inferring reporting times using ERA5 reanalysis.

Objective

• To develop a reliable, readily accessible, and up-to-date observational precipitation dataset specifically for benchmarking AI weather forecasts

Study Configuration

• Spatial Scale: Global (22211 grid cells) with spatial resolution of 0.25°×0.25°
• Temporal Scale: Daily from 2020 to 2025

Methodology and Data

• Models used: Arithmetic mean, Inverse Distance Weighting (IDW), Ordinary Point Kriging (OPK)
• Data sources: Five global station-based observational datasets: GHCN daily, GSOD, GHCNh, IGLD, ISD

Main Results

• The dataset effectively mitigates mismatches in observation reporting times and aligns with the grid resolution of 0.25°×0.25° common in AI weather forecasts.
• Quality control checks improve the median Pearson correlation coefficient (PCC) values from 0.46 to 0.62, 0.54 to 0.71, and 0.33 to 0.47 against IMERG V07 Late, MSWEP V2.8, and GPCC First Guess Daily product, respectively.

Contributions

• The GHCNdgp dataset provides a spatiotemporally consistent observation-based precipitation benchmark that matches the grid resolution of AI weather forecasts.
• The dataset is publicly available on Zenodo in the Network Common Data Form (NetCDF) format, providing essential ground-truth precipitation observations for benchmarking AI weather forecasts.

Funding

• This research was supported by the Southern Marine Science and Engineering Guangdong Laboratory (Zhuhai), School of Civil Engineering, Sun Yat-Sen University, Guangzhou, China.

Citation

@article{Li2026observationbased,
  author = {Li, Qiang and Zhao, Tongtiegang},
  title = {An observation-based daily gridded precipitation dataset for forecast verification},
  journal = {Scientific Data},
  year = {2026},
  doi = {10.1038/s41597-026-08145-8},
  url = {https://doi.org/10.1038/s41597-026-08145-8}
}