Predictability of the 12-month Running Averaged Sunspot Number in the Presence of an 8-month Quasi-periodicity on a Solar Cycle

Abstract

The method of McNish and Lincoln (1949) for the prediction of the 12-month running averaged sunspot number R12 is supposed to generate useful estimates of R12 for periods up to one year ahead. However, it has been noted that for prediction periods beyond about 8 months, the variance of the prediction of R12 (a~12) approaches the variance of the· 'average' solar cycle. Moreover, the variance a~12 for prediction periods greater than about 8 months is also of the order of the variance (a~l) of the observed monthly-mean sunspot data RI. Since the observed sunspot data RI is used to estimate R12, the variance of RI may be used to attach statistical significance to the predictions of R12. Thus, the sunspot number RI2 cannot be usefully predicted more than 8 months ahead, because the variance of the prediction becomes too large (Le. a~12 ~ a~l). However, a quasi-periodicity of about 8 months in RI is observed during the decay phase of solar cycle 21. It is shown in this paper that the variance a~l ought to be doubled in the presence of the 8-month quasi-periodicity of the sunspot cycle. Further, by taking account of this quasi-periodicity, it is possible to make useful predictions of RI2 up to a year (and more) ahead. An application of the RI2 predictions is in forecasting the ionospheric F2 layer critical frequencies a few months ahead.