Zhao et al. (2025) A 1 km daily high-accuracy meteorological dataset of air temperature, atmospheric pressure, relative humidity, and sunshine duration across China (1961–2021)

Identification

Journal: Earth system science data
Year: 2025
Date: 2025-12-17
Authors: Keke Zhao, Denghua Yan, Qin Tianling, Chenhao Li, Dingzhi Peng, Yifan Song
DOI: 10.5194/essd-17-7251-2025

Research Groups

State Key Laboratory of Water Cycle and Water Security, China Institute of Water Resources and Hydropower Research, Beijing 100038, China
College of Water Sciences, Beijing Normal University, Beijing 100875, China

Short Summary

This study presents a 1 km daily high-accuracy meteorological dataset for China (1961–2021), including air temperature, atmospheric pressure, relative humidity, and sunshine duration, generated using a novel hierarchical reconstruction framework that leverages thousands of ground observations and topographic attributes to provide a reliable foundation for climate, hydrological, and ecological studies.

Objective

To develop a 1 km daily high-accuracy meteorological dataset of average, maximum, and minimum air temperature, atmospheric pressure, relative humidity, and sunshine duration across mainland China for the period 1961–2021, to support fine-scale applications in land surface modeling, drought assessment, and water resource management, particularly in data-sparse and topographically complex regions.

Study Configuration

Spatial Scale: Mainland China, 1 km spatial resolution.
Temporal Scale: Daily, 1961–2021 (air temperature); 1961–2019 (atmospheric pressure, relative humidity, sunshine duration).

Methodology and Data

Models used: Multilayer Perceptron (MLP) regression model within a hierarchical progressive reconstruction framework.
Data sources:
- Daily observations from 2345 China Meteorological Administration (CMA) stations (training).
- Independent validation data from 146 stations, including 95 CMA stations, 12 Department of Water Resources (DWR) stations, 31 stations from the National Tibetan Plateau Data Center (TPDC), and 8 Global Surface Summary of Day (GSOD) stations in Taiwan.
- Shuttle Radar Topography Mission (SRTM) Digital Elevation Model (DEM) (version 4.1, 1 km resolution).
- Climate Regionalization Map of China (CMA, 1978).
- For comparison: China Meteorological Forcing Dataset (CMFD 2.0), homogenized daily sunshine duration (SSD) dataset, Himawari AHI-based daily sunshine duration (SD) dataset.

Main Results

The dataset achieved high overall accuracy, validated against 146 independent stations.
Air Temperature (average, maximum, minimum): Median root mean square errors (RMSEs) of 1.16 °C, 1.19 °C, and 1.29 °C; median mean errors (MEs) of -0.04 °C, -0.10 °C, and -0.01 °C; median correlation coefficients (CCs) of 0.99, 0.99, and 0.99, respectively.
Atmospheric Pressure: Median RMSE of 2.65 hPa; median ME of -0.06 hPa; median CC of 0.97.
Relative Humidity: Median RMSE of 6.33 %; median ME of -0.52 %; median CC of 0.90.
Sunshine Duration: Median RMSE of 1.48 h; median ME of 0.05 h; median CC of 0.93.
In high-altitude and topographically complex regions, the reconstructed product demonstrated higher actual accuracy than suggested by station-to-grid validation, with station-based estimates showing greater improvement in RMSE and ME for temperature and pressure.
Compared to CMFD 2.0, the reconstructed dataset exhibited consistently lower median RMSE and ME for average temperature and relative humidity, and substantially smaller errors for atmospheric pressure (median RMSE 3.61 hPa vs 17.14 hPa).
For sunshine duration, the reconstructed dataset achieved temporal accuracy comparable to the homogenized station-based SSD product and spatial accuracy comparable to the high-resolution satellite-based Himawari SD dataset.
Spatial distributions of the reconstructed variables consistently align with China's climatic zonation and physiographic structure, reflecting the influence of latitude, elevation, and oceanic factors.

Contributions

Development of a novel MLP-based hierarchical progressive reconstruction framework that effectively decodes nonlinear relationships between meteorological variables and topographic attributes, enabling the generation of high-accuracy, fine-resolution (1 km) gridded data across complex terrain.
Creation of a comprehensive 1 km daily dataset for China covering six key near-surface meteorological variables (average, maximum, and minimum temperature, atmospheric pressure, relative humidity, and sunshine duration) over an extended period (1961–2021/2019).
Demonstrated improved accuracy and spatial generalizability of the reconstructed dataset compared to existing gridded products (e.g., CMFD 2.0), particularly in challenging, data-sparse, and topographically complex regions like the Tibetan Plateau.
Quantified the underestimation of true product accuracy in station-to-grid validation due to elevation mismatches, especially for elevation-sensitive variables, and showed that station-based estimates using exact coordinates yield more accurate results.
Provides a robust, spatially continuous, temporally complete, and internally consistent dataset that is crucial for advancing regional-scale hydrological, ecological, and climate change studies.

Funding

National Natural Science Foundation of China (grant no. 52130907)
Special Project on Basic Scientific Research Funds of the China Institute of Water Resources and Hydropower Research (grant no. JZ110145B0032025)
Postdoctoral Fellowship Program of the China Postdoctoral Science Foundation (grant no. GZC20233116)
Five Major Excellent Talent Programs of IWHR (grant no. WR0199A012021)

Citation

@article{Zhao20251,
  author = {Zhao, Keke and Yan, Denghua and Tianling, Qin and Li, Chenhao and Peng, Dingzhi and Song, Yifan},
  title = {A 1 km daily high-accuracy meteorological dataset of air temperature, atmospheric pressure, relative humidity, and sunshine duration across China (1961–2021)},
  journal = {Earth system science data},
  year = {2025},
  doi = {10.5194/essd-17-7251-2025},
  url = {https://doi.org/10.5194/essd-17-7251-2025}
}

Original Source: https://doi.org/10.5194/essd-17-7251-2025