The dataset is based on an archive of monthly mean temperatures (CRUTEM 4) provided by more than 5500 weather stations distributed around the world. Each station temperature is converted to an anomaly from the 1961-90 average temperature for that station. As well as the mean anomaly, estimates are made of the uncertainties arising from thermometer accuracy, homogenisation, sampling, large-scale biases such as urbanisation and estimation of regional averages with non-complete global measurement coverage. The data set runs from 1856 and is generally monthly updated.