Lei et al. (2025) Global monthly CMIP6-downscaled high-resolution (1 km) near-surface air temperature dataset (1950–2100)

Identification

• Journal: Scientific Data
• Year: 2025
• Date: 2025-10-28
• Authors: X. G. Lei, Qingyan Meng, Ming Luo, Linlin Zhang, Mijia Yin, Longfei Liu, Qikang Zhao
• DOI: 10.1038/s41597-025-05987-6

Research Groups

• Aerospace Information Research Institute, Chinese Academy of Sciences, Beijing, China
• University of Chinese Academy of Sciences, Beijing, China
• Key Laboratory of Earth Observation of Hainan Province, Hainan Aerospace Information Research Institute, Sanya, China
• School of Geography and Planning, Sun Yat-sen University, Guangzhou, China
• Institute of Geographic Sciences and Natural Resources Research, Chinese Academy of Sciences, Beijing, China
• National Disaster Reduction Center of China, Beijing, China
• State Key Laboratory of Internet of Things for Smart City and Department of Ocean Science and Technology, University of Macau, Macau, China

Short Summary

This study developed MoCHAT, a novel global 1 km monthly near-surface air temperature dataset (1950-2100) by downscaling 16 CMIP6 GCMs using the delta method, providing mean, maximum, and minimum temperatures under three SSP scenarios with high accuracy (mean absolute errors between 1.60 K and 2.38 K).

Objective

• To develop and comprehensively validate MoCHAT, a global, high-resolution (1 km), long-term (1950-2100) monthly near-surface air temperature dataset (including mean, maximum, and minimum temperatures) by downscaling CMIP6 General Circulation Models (GCMs) using the delta method under various Shared Socioeconomic Pathways (SSPs) scenarios.

Study Configuration

• Spatial Scale: Global, 1 km resolution.
• Temporal Scale: Monthly, covering the historical period (1950–2014) and future scenarios (2015–2100) under SSP1-2.6, SSP2-4.5, and SSP5-8.5.

Methodology and Data

• Models used: Delta downscaling method, bilinear interpolation. 16 CMIP6 General Circulation Models (GCMs) from NEX-GDDP-CMIP6 (ACCESS-CM2, ACCESS-ESM1-5, CanESM5, CMCC-ESM2, EC-Earth3, EC-Earth3-Veg-LR, GFDL-ESM4, INM-CM4-8, INM-CM5-0, IPSL-CM6A-LR, MIROC6, MPI-ESM1-2-HR, MPI-ESM1-2-LR, MRI-ESM2-0, NorESM2-LM, NorESM2-MM).
• Data sources:
  • Coarse-resolution GCM data: NASA Earth Exchange Global Daily Downscaled Projections, CMIP6 (NEX-GDDP-CMIP6, 0.25°).
  • Baseline data for downscaling: WorldClim version 2.1 (1 km historical monthly climatology averages, 1970–2000).
  • Validation data: Global Surface Summary of the Day (GSOD) meteorological station observations (daily near-surface air temperatures, 1950–2014).
  • Comparison datasets: CHELSA V2.1 (1 km monthly temperature, 1980–2019), CLIMATE-BCUD (0.1°, East Asia), BCCAQ-CMIP6 (0.25°, global), ERA5-Land.

Main Results

• The MoCHAT dataset provides global monthly mean, maximum, and minimum near-surface air temperatures (tas, tasmax, tasmin) at 1 km resolution from 1950 to 2100, derived from 16 CMIP6 GCMs and their multi-model ensemble (MME).
• Validation against GSOD observations (1950–2014) shows mean absolute errors (MAE) ranging from 1.60 K to 2.38 K and root mean square errors (RMSE) from 2.34 K to 3.27 K across all variables, with overall biases below 2.0 K.
• MoCHAT consistently exhibits higher accuracy (lower MAE and RMSE, stronger correlation) compared to the original NEX-GDDP-CMIP6 data, with the MME performing optimally.
• The probability density functions of MoCHAT closely match GSOD observations, indicating effective preservation of statistical properties.
• Accuracy varies geographically and seasonally: higher errors are observed in North America, Europe, and Asia, and in winter months (DJF), while equatorial regions and summer months (JJA) show better performance. High-elevation regions exhibit biases exceeding 4 K.
• Future projections indicate significant warming, with the high-emission scenario (SSP5-8.5) showing the largest increase rate of 0.53–0.55 K/decade after the 2050s.
• Urban Heat Island Intensity (UHII) in Paris is projected to decrease due to higher warming rates in rural areas compared to urban centers.
• The delta downscaling method, combined with bilinear interpolation, effectively captures fine-scale topographic and climatic variations, outperforming simple bilinear interpolation.

Contributions

• This study presents the first global, monthly, 1 km resolution, long-term (1950-2100) near-surface air temperature dataset derived from CMIP6 GCMs, addressing critical gaps in spatiotemporal resolution.
• The MoCHAT dataset provides unprecedented support for global fine-scale heat risk research, urban climate studies, and the development of urban heat mitigation strategies.
• It offers a comprehensive accuracy assessment, demonstrating superior performance over original GCM outputs and simple interpolation methods, thereby enhancing the reliability of future climate impact assessments.

Funding

• National Natural Science Foundation of China Major Program (grant no. 42192580 and 42192581)
• National Natural Science Foundation of China (grant no. 42201384)
• FY-3 Lot 03 Meteorological Satellite Engineering Ground Application System Ecological Monitoring and Assessment Application Project (Phase I) (grant no. ZQC-R22227)
• Youth Innovation Promotion Association CAS (grant no. 2023139)

Citation

@article{Lei2025Global,
  author = {Lei, X. G. and Meng, Qingyan and Luo, Ming and Zhang, Linlin and Yin, Mijia and Liu, Longfei and Zhao, Qikang},
  title = {Global monthly CMIP6-downscaled high-resolution (1 km) near-surface air temperature dataset (1950–2100)},
  journal = {Scientific Data},
  year = {2025},
  doi = {10.1038/s41597-025-05987-6},
  url = {https://doi.org/10.1038/s41597-025-05987-6}
}