Surface Temperature Anomaly Calculator

A temperature anomaly is a difference, not a temperature: it is the observed value minus the mean of the same location and calendar period over a fixed reference interval. Climatology works in anomalies because absolute temperature is dominated by elevation, latitude and local siting, so two stations 50 km apart can differ by 15 degrees C in absolute terms while their departures from their own long-term means track each other over hundreds of kilometres. That spatial coherence is what makes anomalies averageable onto a grid, and it is why gridded products from NASA GISS, NOAA NCEI, Berkeley Earth and the Met Office HadCRUT series are all published as anomaly fields rather than absolute temperatures. The reference period is a convention and must be stated with any anomaly value. GISTEMP uses 1951 to 1980, HadCRUT reports against 1961 to 1990, NOAA has long used the twentieth-century mean, and the World Meteorological Organization now recommends 1991 to 2020 for operational climate normals. Shifting the baseline shifts every anomaly by a constant but leaves the trend untouched, so a value quoted against a recent baseline will look smaller than the same physical warming quoted against 1850 to 1900. Two cautions apply to the derived warming rate shown here. Dividing a single anomaly by the years since the baseline midpoint assumes a linear trend through that midpoint, which interannual variability of roughly 0.1 to 0.2 degrees C from El Nino and volcanic eruptions can easily distort, and extrapolating that rate to 2050 or 2100 is arithmetic rather than a climate projection.

The idea of averaging station records into a single global figure dates to the nineteenth century, when Wladimir Koppen compiled temperature series from several hundred stations in the 1870s and early 1880s to look for worldwide fluctuations. The first influential attempt to link such a compilation to carbon dioxide came from Guy Stewart Callendar, a British steam engineer, who assembled records from around 200 stations and argued in a 1938 paper to the Royal Meteorological Society that world temperatures had risen over the preceding half century and that fossil fuel emissions were responsible. Modern anomaly datasets were built in the 1980s once digital archives and computing made global gridding feasible. James Hansen and Sergej Lebedeff at the NASA Goddard Institute for Space Studies published their global surface air temperature analysis in 1987, establishing the station-anomaly-then-interpolate method that GISTEMP still uses, while Phil Jones and colleagues at the Climatic Research Unit produced parallel land and marine compilations that became the HadCRUT series with the Met Office Hadley Centre. Persistent methodological work has since focused on homogenization of station discontinuities, correction of the wartime shift in sea surface measurement from canvas buckets to engine intakes, and coverage of the sparsely sampled Arctic, addressed by kriging in Berkeley Earth and in HadCRUT5, described by Morice and colleagues in 2021.