HadCRUT5¶

HadCRUT5 — Met Office Hadley Centre / Climatic Research Unit global surface temperature anomalies

Available

At a glance¶

Note on period: As checked on 2026-10-12, the HadCRUT.5.2.0.0 annual summary files include a (partial) 2026 value, and the accompanying CRUTEM.5.2.0.0 release notes state "2026 based on data for the first seven months of the year only", so data appear to run through July 2026. The last row of the monthly files was not checked directly.

Overview¶

A gridded dataset of global historical surface temperature anomalies from January 1850, produced jointly by the Met Office Hadley Centre and the Climatic Research Unit at the University of East Anglia. It blends the CRUTEM5 land air temperature and HadSST4 sea surface temperature datasets on a 5° grid, with anomalies relative to 1961–1990. It comes in a non-infilled form, holding only cells with observations, and a statistically infilled analysis; both are presented as an ensemble of 200 realisations that sample the uncertainty. It is one of the main global temperature datasets used in IPCC and WMO assessments.

Available variables¶

Products as listed on the official download page. CSV column names are as they appear in the files. Variable names inside the gridded NetCDF files were not checked for this page and are not listed (inspect them with ncdump -h after download).

instantaneous mean accumulated maximum minimum static — instantaneous: value at the time stamp · mean/accumulated/maximum/minimum: over the indicated interval · static: does not change with time

Gridded data (analysis and non-infilled)¶

Summary series (global (NH+SH)/2, Northern and Southern Hemisphere)¶

The full variable list is in the official documentation.

How to get the data¶

• Met Office Hadley Centre download page (HadCRUT.5.2.0.0) — HTTPS (web / wget / curl) No login needed
  • Separate sections for the analysis and the non-infilled data, each with summary, ensemble and uncertainty component series (CSV and NetCDF, monthly and annual), gridded ensemble members (NetCDF zip files), and ensemble mean and weights files. File names include the version (HadCRUT.5.2.0.0), so URLs change with each new version.
• Previous versions (5.0.0.0 to 5.1.0.0) — HTTPS No login needed
  • Use these to reproduce published results. Each version has release notes describing what changed.
• CEDA Archive (HadCRUT.5.0.0.0) — CEDA web download / OPeNDAP
  • Archive of the first version (1850–2018). Get current data from the Met Office pages.

Example code¶

Reads the analysis global monthly summary series (CSV) and re-expresses it relative to 1850–1900. Update the version number in the URL to the current version.

import pandas as pd

ver = "HadCRUT.5.2.0.0"
url = (f"https://www.metoffice.gov.uk/hadobs/hadcrut5/data/{ver}/analysis/diagnostics/"
       f"{ver}.analysis.summary_series.global.monthly.csv")
df = pd.read_csv(url)
df["Time"] = pd.PeriodIndex(df["Time"], freq="M")
s = df.set_index("Time")["Anomaly (deg C)"]           # °C, relative to 1961-1990

# Subtract the 1850-1900 mean to get anomalies relative to pre-industrial
s_pi = s - s["1850":"1900"].mean()
print(s_pi.tail(12).round(2))

Practical tips¶

• The analysis and non-infilled versions give different global means. The non-infilled version leaves out empty cells such as the sparsely observed Arctic, so do not mix or splice the two.
• Anomalies are relative to 1961–1990, while GISTEMP and Berkeley Earth use 1951–1980. Re-baseline to a common period before comparing datasets.
• To carry the uncertainty properly, compute your statistic for all 200 members and look at the spread rather than using the ensemble mean alone. For the non-infilled data, also add the measurement and sampling errors held outside the ensemble (uncorrelated and error_covariance files).
• The Open Government Licence requires attribution. Include the version number, the download date and "© British Crown Copyright, Met Office" (see the citation guidance below).
• The HadCRUT.5.2.0.0 release notes say the ocean input changed "from HadSST.4.1.0.0 to HadSST.4.1.1.0" but link to HadSST.4.2.0.0 documents. If the exact HadSST version matters, ask the Met Office.
• As of October 2026 the CRU (UEA) data page still points to 5.0.2.0 (data to May 2024). Download the latest version from the Met Office pages.
• Land station data are updated country by country (in 5.2.0.0, for example, Chinese and US stations were replaced), so regional averages can shift noticeably between versions.

Notices and change history¶

References for your paper¶

Dataset citation¶

Attribution required by the Met Office (Open Government Licence)

@misc{hadcrut5_dataset,
  author       = {{Met Office Hadley Centre} and {Climatic Research Unit, University of East Anglia}},
  title        = {{HadCRUT}.5.2.0.0: Global surface temperature anomalies},
  year         = {2026},
  howpublished = {\url{https://www.metoffice.gov.uk/hadobs/hadcrut5/}},
  note         = {Open Government Licence v3. Accessed DD MMM YYYY}
}

Reference papers¶

HadCRUT5 paper (Met Office recommended citation) — DOI 10.1029/2019JD032361

@article{morice2021hadcrut5,
  author  = {Morice, C. P. and Kennedy, J. J. and Rayner, N. A. and Winn, J. P. and Hogan, E. and Killick, R. E. and Dunn, R. J. H. and Osborn, T. J. and Jones, P. D. and Simpson, I. R.},
  title   = {An Updated Assessment of Near-Surface Temperature Change From 1850: The {HadCRUT5} Data Set},
  journal = {Journal of Geophysical Research: Atmospheres},
  volume  = {126},
  number  = {3},
  pages   = {e2019JD032361},
  year    = {2021},
  doi     = {10.1029/2019JD032361}
}

Acknowledgements¶

The Met Office terms ask you to cite Morice et al. (2021) and to include the attribution statement above (with version number, download date and year of first publication). We found no separate dataset DOI. If you use the component datasets on their own, also cite CRUTEM5 (Osborn et al. 2021) and HadSST4 (Kennedy et al. 2019).

Information on the datasets used in this study was compiled with the help of the Korean Climate Data Navigator (KCDN) (Sim, 2026; https://doi.org/10.5281/zenodo.23300923).

Links¶

• HadCRUT5 (Met Office Hadley Centre)
• HadCRUT.5.2.0.0 release notes
• HadCRUT5 terms and conditions
• CRUTEM5 (land input)
• HadSST4 (ocean input)
• CRU temperature data page (UEA)
• Met Office climate dashboard — global temperature