Periodicities in solar wind-magnetosphere coupling functions and geomagnetic activity during the past solar cycles

Abstract

In this paper, we study the solar-terrestrial relation through the wavelet analysis. We report periodicities common between multiple solar wind coupling functions and geomagnetic indices during five solar cycles and also and the strength of this correspondence. The $\mathit{Dst}$ (found to be most predictable in Newell et al., J. Geophys. Res. Space Phys. 112(A1):A01206, 2007) and $\mathit{AL}$ (least predictable in Newell et al., J. Geophys. Res. Space Phys. 112(A1):A01206, 2007) indices are used for this purpose. During the years 1966–2016 (which includes five solar cycles 20, 21, 22, 23, and 24), prominent periodicities $\leq720$ days with power above 95% confidence level were found to occur around 27, 182, 385, and 648 days in the $\mathit{Dst}$ index while those in the $\mathit{AL}$ index were found in bands around 27, 187, and 472 days. Ten solar wind coupling functions were then used to find periodicities common with the indices. All the coupling functions had significant power in bands centered around 27, 280, and 648 days while powers in fluctuations around 182, 385, and 472 days were only found in some coupling functions. All the drivers and their variants had power above the significant level in the 280–288 days band, which was absent in the $\mathit{Dst}$ and $\mathit{AL}$ indices. The normalized scale averaged spectral power around the common periods in the coupling functions and the indices indicated that the coupling functions most correlated with the $\mathit{Dst}$ index were the Newell (27 and 385 days), Wygant (182 days), and Scurry-Russell and Boynton (648 days) functions. An absence of common power between the coupling functions and the $\mathit{Dst}$ index around the annual periodicity was noted during the even solar cycles. A similar analysis for the $\mathit{AL}$ index indicated that Newell (27 days), Rectified (187 days), and Boynton (472 days) were the most correlated functions. It was also found that the correlation numbers were relatively weaker for the $\mathit{AL}$ index, specially for the 187 day periodicity. It is concluded that as the two indices respond to solar wind forcing with varying levels of strength at various prominent scales and the coupling function used, the response might be dependent on the scale (days or months or years) of interest at which the solar wind driving is to be predicted.