Determination of solar cycle length variations using the continuous wavelet transform

The length of the sunspot cycle determined by FriisChristensen & Lassen (1991) correlates well with indicators of terrestrial climate, but has been criticized as being subjective. In the present paper we present a more objective and general cycle-length determination. Objectivity is achieved by using the continuous wavelet transform based on Morlet wavelets and carrying out a careful error analysis. Greater generality comes from the application of this technique to different records of solar activity, e.g. sunspot number, sunspot area, plage area or 10Be records. The use of different indicators allows us to track cycle length variations back to the 15th century. All activity indicators give cycle length records which agree with each other within the error bars, whereby the signal due to the solar cycle is weaker within 10Be than in the other indicators. In addition, all records exhibit cycle length variations which are, within the error bars, in accordance with the record originally proposed by Friis-Christensen & Lassen (1991). In the 16th century, however, the 10Be record suggests a much longer cycle than the auroral record used by Friis-Christensen & Lassen. Also, the presence of a distinct 11-year cycle in the 10Be record during the Maunder Minimum is confirmed. By combining the results from all the indicators a composite of the solar cycle length is constructed, which we expect to be more reliable than the length derived from individual records.