Secular variation of recurrent geomagnetic disturbance

Analysis is made of the secular variation of the occurrence and recurrence statistics of magnetic storms and lower-levels of geomagnetic activity using K-index and aa-index data obtained from ground-based magnetic observatories, 1868.0-2009.0, solar cycles 11-23. Methods include parametric measures of correlation, autocorrelation, and conditional probability, and to accommodate calibration differences from data collected independently at different observatories, nonparametric measures of trend and correlation are used. Consistent signals can be interpreted in terms of solar-terrestrial interaction: (1) In rough correspondence to a general increase in sunspot number, over the past 142 years the occurrence and 27.0-d-recurrence of geomagnetic disturbance have both increased, but these changes are not simple linear trends. (2) The general increase in geomagnetic disturbance comes despite a slightly proportional decrease in 27.0-d recurrent geomagnetic disturbance. From this we infer that the role of high-speed streams as a driver of geomagnetic activity has slightly decreased in comparison to coronal-mass ejections. (3) Occasionally, geomagnetic disturbance shows 13.5-d recurrence, and there is even a hint of 9.0-d recurrence. From this we infer that the polarity of the heliodynamo can reverse via a variety of transitional routes, some of which exhibit more poloidal dipole symmetry than others. (4) The statistical relationship over the past 142 years between sunspot number, geomagnetic disturbance and geomagnetic recurrence is nonstationary. From this we infer that it is extremely challenging to backward extrapolate correlations involving modern data to understand the past historical evolution of the Sun. Likewise, it is extremely challenging to predict the number of future sunspot numbers, but there might be an opportunity to use nonstationary solar-terrestrial variables to test theories of coupling between the solar wind and the magnetosphere.