Annual fractions of high‐speed streams from principal component analysis of local geomagnetic activity

Abstract

We study the latitudinal distribution of geomagnetic activity in 1966–2009 with local geomagnetic activity indices at 26 magnetic observatories. Using the principal component analysis method we find that more than 97% of the variance in annually averaged geomagnetic activity can be described by the two first principal components. The first component describes the evolution of the global geomagnetic activity, and has excellent correlation with, e.g., the Kp/Ap index. The second component describes the leading pattern by which the latitudinal distribution of geomagnetic activity deviates from the global average. We show that the second component is highly correlated with the relative (annual) fraction of high‐speed streams (HSS) in solar wind. The latitudinal distribution of the second mode has a high maximum at auroral latitudes, a local minimum at subauroral latitudes and a low maximum at midlatitudes. We show that this distribution is related to the difference in the average location and intensity between substorms related to coronal mass ejections (CMEs) and HSSs. This paper demonstrates a new way to extract useful, quantitative information about the solar wind from local indices of geomagnetic activity over a latitudinally extensive network.

A new method to estimate annual solar wind parameters and contributions of different solar wind structures to geomagnetic activity

“…In this paper we have studied the spatiotemporal evolution of geomagnetic activity in 1966–2011 using local A h and IHV indices of 26 stations covering a wide range of latitudes. We analyzed the indices using the principal component analysis method and confirmed that our recent results for the A h indices [ Holappa et al , ] also hold for IHV indices, i.e., that the first PC describes global average geomagnetic activity and the second PC the deviations from the global average caused by high‐speed streams.…”

“…The respective EOF1(Ah) and EOF1(IHV) depicted in Figure describe the latitudinal modes associated with the PC1s. As we found earlier [ Holappa et al , ], EOF1(Ah) is almost flat (independent of latitude), meaning that all stations contribute with roughly equal weights to PC1. Hence, the PC1(Ah) is very closely proportional to the average of the 26 A h s indices.…”

“…As noted earlier [ Holappa et al , ] PC2(Ah) is very highly correlated (cc =0.82; p =4.6·10 −12 ) with the annual time fraction of high‐speed streams in solar wind. This can also be seen in Figure which shows the annual fraction of HSSs in solar wind according to the classification of solar wind into three flow types [ Richardson and Cane , ].…”

“…Recently, we used the principal component analysis (PCA) method to extract information on the solar wind drivers of annually averaged geomagnetic activity using a set of local A h indices [ Holappa et al , ]. We found that the first principal component (PC1) represents the global average of the A h indices and correlates almost perfectly with the A p index and that the second principal component (PC2) highly correlates with the annual fraction of high‐speed streams in the solar wind.…”

Long‐term decrease in the response of midlatitude stations to high‐speed solar wind streams in 1914–2000

“…On the other hand, when disturbances are directly driven by CMEs around solar maxima, NGK, due to its weaker dependence on B , registers suppressed AhK Z . This is in agreement with recent results of Holappa et al [], where it was demonstrated that the fraction of HSS (CME)‐related activity in annual geomagnetic activity depicts a local maximum (minimum, respectively) at the specific range of midlatitudes where NGK is situated. The different dependence to HSSs versus CMEs was also held accountable for the dominant fraction of observed differences in geomagnetic activity between the geomagnetic stations.…”