We assess the random and systematic uncertainties affecting the central England temperature (CET) record since 1878 on daily, monthly and annual time scales. The largest contribution to uncertainty in CET on all these time scales arises from areal sampling, followed for annual and monthly CET by thermometer calibration. For the daily series, random thermometer precision and screen errors are the second largest source of uncertainty. Annual CETs are least uncertain, whereas daily CETs are most uncertain. Despite the uncertainties in annual mean CET, the trend of 0.077 °C per decade since 1900 is significant at the 1% level.
In an additional investigation, we detect biases in the published series of central England maximum and minimum temperatures, and implement systematic adjustments of up to ±0.2 °C to the values up to 1921 and up to ±0.1 °C to the values since 1980. These adjustments are of opposite sign in maximum and minimum temperature, so they do not affect mean CET, but they improve the homogeneity of the diurnal temperature range, which then shows little trend before 1980 and a reduced rising trend thereafter. The uncertainties in maximum and minimum temperature make the data inadequate for the task of establishing the magnitude of the recent increase of diurnal range. © Crown Copyright 2005. Reproduced with the permission of Her Majesty's Stationery Office. Published by John Wiley & Sons, Ltd.