The Maunder Minimum

Posted by James Watson on 03 Sep, 2016

A pdf version of this post available here.

The ‘solar cycle’ describes the cyclic wax and wane in solar activity over a (nominal) 11 year period. A common proxy for expressing solar activity is the ‘Sun Spot Number’ (SSN) a count of the number of dark spots observed on the sun’s surface. The SSN has been found to correlate with solar activity, the SSN rising in times of increased solar activity favouring enhanced HF conditions. Larger sun spots are visible to the naked eye and have been observed as long ago as 4BC (Eddy, 1976) with useful records dating back for around 400 years, providing a wealth of historical data (although it is debatable if the expression ’historical’ may be applied to a period of 400 years in relation to the life of a star).

Cycles are subject to fluctuation in terms of both size and duration. Peak values have been observed to range from 68 in 1918 to 225 in 1959 (SILSO, 2016) and cycle durations observed from 9–14 years (Hathaway et al., 1994)

The ‘Maunder Minimum’ refers to a 70 year period between the years 1645– 1715 noted for extremely low levels of solar activity; not a single sun spot was observed in the period 1672–1704 (Eddy, 1976). Wittmann (1978) examines evidence from the years immediately prior to the Maunder Minimum and con- cludes that sun spot sightings leading to the event follow an 11-year solar cycle. The Maunder Minimum therefore represents an abrupt cessation of activity in an otherwise regular phenomena; a troubling observation to those seeking to characterise solar behaviour in terms of a single cyclical activity, implying as it does a mechanism capable of stopping and restarting of its own accord. This may have contributed to some initial resistance from sceptics who argued that the apparent period of low activity was simply due to the paucity of information and / or the reliability of any surviving observations from the time period. Re- cent studies citing evidence from 88 observers concur with Maunder’s position that the phenomena did indeed exist Hoyt and Schatten (1996).

Eddy (1976) speculates that fluctuations in the solar cycle are driven by larger underlying mechanisms, employing an analogy familiar to radio engineers; that of the 11-year cycle serving as a ‘carrier’, modulated by a longer term behaviour. Wittmann (1978) presents a statistical analysis to support this idea, suggesting the presence of multiple cycles with periods of 92, 55 and 10 years, albeit with no more than 25% of the influence of the established 11 year cycle. Multiple cycles of differing periodicity operating simultaneously would continuously shift phase in a manner that could produce the observed effects. If the Maunder Minimum is then accepted as the result of natural phenomena operating within the sun then one could argue that it is only reasonable to accept that it could happen again and if so, is it possible to predict a recurrence?

Writing prior to the start of Cycle 24, Pesnell (2008) surveys more than 50 predictions for the cycle, categorising them into the following types;

The maximum values from individual predictions ranged from 40 to 185, a range that encompasses virtually all activity since 1705. In the event, Cycle 24 has proven to be similar in shape and size to Cycle 16, peaking at an annual sunspot number of about 80.

As the cycle draws to a close, Pesnell (2016) reviews the accuracy of the surveyed predictions, observing that the overall average value of 106 is some way above the observed value (See Note 1 below). Given that the majority of the predictions are based upon past observations this may not be entirely surprising for a period following the ‘Grand Maximum’ of the 5 previous cycles. It says much about the current state of solar forecasting that as recently as 2006, predictions of cycle 24 “going to be one of the most intense cycles since record-keeping began almost 400 years ago” were being made NASA (2009).

Writing in this month’s ‘The Spectrum Monitor’, Reitz (2016) speculates that the recent trend in depressed solar cycle maxima is a precursor to a repeat of the Maunder Minimum, a view echoed by Shepherd et al. (2014) who forecast low solar activity throughout cycles 25-26, approaching levels observed in the Maunder Minimum. The publication of Shepherd et al.’s (2014) work led to a flurry of media activity, debating possible effects on the global climate (a topic not actually addressed by Shepherd et al.) and detracting somewhat from the main aspect of the paper; a modified model for solar prediction. Shepherd et al. (2014) propose a revised solar model, incorporating pairs of magnetic waves originating at two layers of the sun’s interior. A model was constructed and ‘trained’ on historical data using the Eureqa software application to determine underlying laws. In demonstrating the validity of the revised model, the authors compare measured and theoretical values with encouraging accuracy. The slight lift in values observed in 2014 is also successfully predicted in the paper (published prior to the event). The model forecasts a continued depression in solar activity in cycles 25 and 26, to levels of 80% and 40% of the maximum levels observed during the current cycle. This would correspond to maximum SSN values of 91 and 46 for cycles 25 and 26 respectively (assuming the maximum monthly smoothed value observed during cycle 24 (June 2014) of 114.1 SILSO (2016).

When assessing the effect that such values have on HF communications, it is useful to place these values into the context of previous cycles using data pro- vided by SILSO. Over the past 100 years, 53% of the smoothed monthly sun spot values are lower than or equal to the predicted peak of cycle 25 and 32% are lower than the predicted peak of Cycle 26. Operators will experience in- creased difficulty in establishing robust links but no more so than during the extended minima between cycles 23 and 24 in the years 2007-2009, during which time monthly smoothed sun spot numbers were measured in single digits. The advent of a period of low solar activity should not therefore lead to a demise in the importance of HF radio as a communications media.


  1. In July 2015, the SIDC adopted an revised scale in which values are 40-70% larger than the previous version. For consistency, Pesnell (2016) adopts the older scale used in the original survey in the follow-up review.


Eddy, J. A. (1976, June). The Maunder Minimum. Science, 192:1189–1202. doi: 10.1126/science.192.4245.1189.

Hathaway, D. H., Wilson, R. M., and Reichmann, E. J. (1994). The shape of the sunspot cycle. Solar Physics, 151(1):177–190.

Hoyt, D. V. and Schatten, K. H. (1996). How well was the sun observed during the maunder minimum? Solar Physics, 165(1):181–192.

NASA (2009). Scientists predict big solar cycle. NASA Science: Science News, [Online] Available from: at-nasa/2006/21dec_cycle24/ (Accessed: 23 August 2016).

Pesnell, W. D. (2008). Predictions of solar cycle 24. Solar Physics, 252(1): 209–220.

Pesnell, W. D. (2016). Predictions of solar cycle 24: How are we doing? Space Weather, 14(1):10–21.

Reitz, K. (2016, September). Radio 101: Solar Cycle 25 -or- How I Learned to Stop Worrying and Love the Mauder Minimum. The Spectrum Monitor, 3: 61–64.

Shepherd, S. J., Zharkov, S. I., and Zharkova, V. V. (2014). Prediction of solar activity from solar background magnetic field variations in cycles 21-23. The Astrophysical Journal, 795(1):46, [Online] Available from: http://stacks.

SILSO (2016, Sep). Yearly mean total sunspot number, [Online] Available from: (Accessed: 23rd Au- gust 2016).

Wittmann, A. (1978). The sunspot cycle before the maunder minimum. As- tronomy and Astrophysics, 66:93.