Evidence for solar‐cycle forcing and secular variation in the Armagh Observatory temperature record (1844–1992)

Abstract

A prominent feature of previous long‐term temperature studies has been the appearance of warming since the 1880s, this often being taken as evidence for anthropogenic‐induced global warming. In this investigation, the long‐term, annual, mean temperature record (1844–1992) of the Armagh Observatory (Armagh, North Ireland), a set of temperature data based on maximum and minimum thermometers that predates the 1880s and correlates well with northern hemispheric and global standards, is examined for evidence of systematic variation, in particular, as related to solar‐cycle forcing and secular variation. Indeed, both appear to be embedded within the Armagh data. Removal of these effects, each contributing about 8% to the overall reduction in variance, yields residuals that are randomly distributed. Application of the 10‐year moving average to the residuals, furthermore, strongly suggests that the behavior of the residuals is episodic, inferring that (for extended periods of time) temperatures at Armagh sometimes were warmer or cooler (than expected), while at other times they were stable. Comparison of cyclic averages of annual mean temperatures against the lengths of the associated Hale cycles (i.e., the length of two, sequentially numbered, even‐odd sunspot cycle pairs) strongly suggests that the temperatures correlate inversely (r = −0.886 at <2% level of significance) against the length of the associated Hale cycle. Because sunspot cycle 22 ended in 1996, the present Hale cycle probably will be shorter than average, implying that temperatures at Armagh over this Hale cycle will be warmer (about 9.31±0.23°C at the 90% confidence level) than average ( = 9.00°C).