Lefèvre, Laure ; Dudok de With, Thierry ; Clette, Frédéric

Poster presented at Space Climate 6 Levi Finland on 2016-04-04

Abstract: The International Sunspot Number recently underwent an extensive makeover. In this context, a first estimation of the error bars related to this well-known measurement of solar activity was added. Uncertainties can belong to different categories. Accuracy (differences in absolute calibration and instrument bias) and stability ,(associated with long-term drifts), are taken care of by the k-factors, and are not in the scope of this study. We focus on the effect of random errors, which are usually associated with precision and repeatability, namely how close independent measurements are when made under identical conditions. These random errors can be inferred from a time series analysis of sunspot number records, we call these time domain errors, and they are mostly related to the precision. A second type of error arises when comparing different observers, and considering their dispersion. Here we study their comparative behaviours to understand where they stem from and how to infer them. Around solar maximum, it seems it is not the Sun that is dominating the uncertainties, but observational errors. At first glance we could say that for high sunspot numbers, the grouping of different active regions between observers might create this bias. As it turns out, the number of spots presents the most similarities with the level of uncertainties: the ability of observers to count all the spots is put to the test during high activity phases.