Een verband tussen zonnevlekken en extreme winters

A relation between solar activity and extreme winters The well‐known II‐years period in solar activity is eliminated by computing moving averages of relative sunspot numbers over II‐year intervals. The averages thus obtained reveal a ‘long period’ probably of about 89 years. By subtracting the averages from the original figures reduced sunspot numbers are obtained which represent the normal II‐years fluctuations after elimination of the long period. Standard deviations of winter temperatures have been computed in a similar fashion for moving II‐year intervals, and these standard deviations have been plotted against the average sunspot numbers in figs. 1 and 2, crosses referring to the years before 1831 and dots to the years after. Both at De Bilt and at Prague the observations from 1831 onwards exhibit a pronounced positive correlation, with correlation coefficients of +0.82 and +0.74 respectively. No such correlation can be detected in the earlier observations, which the author is, however, inclined to ascribe to the inaccuracy and incompleteness of the data collected in these early periods. As shown by fig. 3 no correlation does exist between the temperature deviations and the reduced sunspot numbers. It must be concluded that extreme winters occur during the peak solar activities of the long solar period; the normal II‐years period is of no consequence. From a statistical point of view these results are noteworthy: the standard deviation shows systematic variations which are doubtless of physical origin. Statisticians are advocated not to lose sight of the physical background when applying statistical methods to physical phenomena.