The Discontinuity Circa 1885 in the Group Sunspot Number

On average, the international sunspot number (\(R_{\mathrm{I}}\)) is 44 % higher than the group sunspot number (\(R_{\mathrm{G}}\)) from 1885 to the beginning of the \(R_{\mathrm{I}}\) series in 1700. This is the principal difference between \(R_{\mathrm{I}}\) and \(R_{\mathrm{G}}\). Here we show that this difference is primarily due to an inhomogeneity in the Royal Greenwich Observatory (RGO) record of sunspot groups (1874 – 1976) used to derive observer normalization factors (called \(k\)-factors) for \(R_{\mathrm{G}}\). Specifically, annual RGO group counts increase relative to those of Wolfer and other long-term observers from 1876 – 1915. A secondary contributing cause is that the \(k\)-factors for observers who began observing before 1884 and overlapped with RGO for any years during 1874 – 1883 were not based on direct comparison with RGO but were calculated using one or more intermediary or additional observers. We introduce \(R_{\mathrm{GC}}\) by rectifying the RGO group counts from 1874 – 1915 and basing \(k\)-factors on direct comparison with RGO across the 1885 discontinuity, which brings the \(R_{\mathrm{G}}\) and \(R_{\mathrm{I}}\) series into reasonable agreement for the 1841 – 1885 interval (after correcting \(R_{\mathrm{I}}\) for an inhomogeneity from 1849 – 1867 (to give \(R_{\mathrm{IC}}\))). Comparison with an independently derived backbone-based reconstruction of \(R_{\mathrm{G}}\) (\(R_{\mathrm{BB}}\)) indicates that \(R_{\mathrm{GC}}\) over-corrects \(R_{\mathrm{BB}}\) by 4 % on average from 1841 – 1925. Our analysis suggests that the maxima of Cycles 10 (in 1860), 12 (1883/1884), and 13 (1893) in the \(R_{\mathrm{IC}}\) series are too low by ≈ 10 %.