Three Centuries of Sunspot Group Numbers with Monthly Resolution

Abstract:

I construct four overlapping, intercalibrated 'backbone' composite series [based on primary observers Staudach, Schwabe, Spōrer, Wolfer, and Greenwich (RGO)] covering the three-century interval 1720-2020 of the Sunspot Group Number with monthly resolution. Of note are the following properties:

• There is no significant long-term trend (linear R² = 0.0045), thus no Modern Grand Maximum
• In particular, Cycle 11 (max. in 1870) is on par with the high cycles in the 20th century
• The regressions between backbones are linear and and even proportional (offsets not statistically different from zero)
• The regressions between individual observers within each backbone are linear with offsets not statistically different from zero
• Because each observer within a backbone is compared directly with the primary observer without any intermediate observers there is no 'daisy-chaining' and thus no propagation of errors
• Because the backbones overlap, they can be compared directly with the previous and following backbones, thus minimizing inter-backbone daisy-chaining (at most two links)
• The new composite is statistically indistinguishable from the published Svalgaard & Schatten [2016] composite (yearly) series
• The new composite is compatible with all known well-constrained proxies and solar activity indices (in particular, the revised Version 2 sunspot number)
• The new composite supersedes those reconstructions of Group Numbers that show significantly lower activity in the 19th century (especially during Cycle 11)
• Within each backbone, the carefully normalized series is no better than the simple average [with no normalization at all] of all observers at the time.
• The secular increase (from one backbone to the next) in archived Group Numbers ["the data"] is probably due to evolving technology and understanding of what makes a group, rather than to errors and mistakes committed by researchers attempting to reconstruct the long-term series of the Group Number
• And that we, therefore, have several 'populations' of group numbers [roughly corresponding to the backbones] with different statistical properties evolving with time
• So that the true evolution of solar activity can only be validated by agreement with other manifestations of said activity (often derisively called ‘proxies’) of which there are many ("everything must agree")