What is the Central England Temperature?
The Central England Temperature (CET) series is a record of temperature across a small part of England. It goes back to 1659, which makes it the longest instrumental temperature record in the world.
The area it represents
CET is not a single thermometer. It represents a roughly triangular area of lowland England with its corners near Lancashire, London and Bristol, which takes in much of the Midlands. Over the years, observations from a series of weather stations inside or near this triangle have been combined. The aim is a single, consistent value for each day and month.
A short history
- 1953 Climatologist Gordon Manley publishes a monthly mean temperature series for central England covering 1698–1952. He built it from early instrumental records and observers' diaries.
- 1974 Manley publishes an extended and revised monthly series covering 1659–1973 (Quarterly Journal of the Royal Meteorological Society). This is still the basis of the early record.
- 1992 Parker, Legg & Folland (Met Office) publish a daily mean series from 1772 and bring the record up to date using modern stations. The Met Office Hadley Centre has maintained the series as HadCET since then.
- 2004–06 Station changes. Stonyhurst (Lancashire) replaces the combination of Squires Gate (Blackpool) and Ringway (Manchester) from November 2004. Pershore College replaces Malvern in January 2006.
- 2005 Parker & Horton add daily maximum and minimum series from 1878, with an assessment of their uncertainties.
- 2022 HadCET v2.0 reduces the number of station changes in the record. Rothamsted is used instead of Cambridge for 1878–1930, and the Pershore switch is moved to November 2004. The update also improves traceability and version control.
- 2024 v2.1 corrects minor clipping errors in station mean temperatures up to the end of 2020. It is described in Legg et al. (2025), Geoscience Data Journal.
- 2026 v2.1.1.0 (March) extends the version-controlled dataset through 2025.
How it has been calculated over the years
1659–1771: monthly values from historical sources. No single long, well-exposed record exists for this period. Manley therefore spliced together overlapping records from many observers, adjusting each to a common standard. The earliest thermometers were often kept indoors or in unheated rooms, and some gaps were filled from weather diaries. The precision of the published values reflects this. You can see it in the monthly data file itself:
| Years | Monthly values given to the nearest… |
|---|---|
| 1659–1670 | 1 °C |
| 1671–1698 | 0.5 °C |
| 1699–1706 | 0.1 °C |
| 1707–1721 | 0.5 °C |
| 1722 onwards | 0.1 °C |
1772 onwards: a daily series. Parker et al. (1992) built daily mean temperatures from a succession of station records, made consistent with Manley's monthly means. Early examples include Kew/Syon House and Lyndon Hall in Rutland, which covered 1777–1789. From 1878 the daily maximum and minimum series follow the same principle (Parker & Horton, 2005).
Modern period: a small set of stations averaged together. Since 1974 the series has been calculated from a few stations spread across the triangle, adjusted for differences between sites. Today these are Rothamsted (Hertfordshire), Pershore College (Worcestershire) and Stonyhurst (Lancashire). Whenever a station is replaced, the record is homogenised so the switch does not introduce a step change.
Allowing for urban warming. Since 1974 a correction is applied to mean temperatures to remove the warming effect of growing towns around the stations. It is currently −0.2 °C.
Live updates. Provisional daily values are added each day. The monthly value is normally available by the 3rd of the following month. Recent data are provisional and can be revised. Because the updates are automated, the files can occasionally lag by a day or two.
Monthly, seasonal and annual values. In the published files, a monthly value is the mean of that month's daily values, and a seasonal value is the mean of its three monthly values. The seasons are winter (Dec–Feb, labelled by the year of its January), spring (Mar–May), summer (Jun–Aug) and autumn (Sep–Nov). The annual value is the mean of all the year's daily values. The dashboard uses the same rules to work out values for the period still in progress.
How reliable is it?
Uncertainty is largest in the earliest decades, when instruments, their exposure and observing practices differed a lot from today. From the late 19th century onwards the record is considered robust. Parker & Horton (2005) quantify the uncertainty in the series since 1878. Small differences between years, especially a few hundredths of a degree in seasonal or annual values, are well within the uncertainty. Close rankings should be read with that in mind.
Why it matters
CET covers more than 360 years. That makes it a key reference for natural climate variability in the pre-industrial era, for example the cold decades of the late 17th century, and for the warming of recent decades. It is also one of the most widely quoted climate records in the UK.
Further reading
- Met Office Hadley Centre: HadCET dataset home page
- HadCET release notes and version history
- Data download page
- Parker, Legg & Folland (1992), A new daily Central England Temperature series, 1772–1991. Int. J. Climatol.
- Parker & Horton (2005), Uncertainties in the Central England Temperature series since 1878. Int. J. Climatol.
- Manley, G. (1953), The mean temperature of central England, 1698–1952; and (1974), Central England temperatures: monthly means 1659 to 1973. Q. J. R. Meteorol. Soc.
- Legg, T. et al. (2025), An update to the Central England Temperature series: HadCET v2.1. Geoscience Data Journal.
Data: Met Office Hadley Centre HadCET. Facts on this page are taken from the HadCET website and release notes. See those for full details of station history and methods.