Geomagnetic Effects of Corotating Interaction Regions

Abstract

We present an analysis of the geoeffectiveness of corotating interaction regions (CIRs), employing the data recorded from 25 January to 5 May 2005 and throughout 2008. These two intervals in the declining phase of Solar Cycle 23 are characterised by a particularly low number of interplanetary coronal mass ejections (ICMEs). We study in detail how four geomagnetic-activity parameters (the Dst, Ap, and AE indices, as well as the Dst time derivative, $\mathrm{dDst}/\mathrm{d}t$) are related to three CIR-related solar wind parameters (flow speed, $V$, magnetic field, $B$, and the convective electric field based on the southward Geocentric solar magnetospheric (GSM) magnetic field component, $\mathit{VB}_{s}$) on a three-hour time resolution. In addition, we quantify statistical relationships between the mentioned geomagnetic indices. It is found that Dst is correlated best to $V$, with a correlation coefficient of $\mathrm{cc}\approx0.6$, whereas there is no correlation between $\mathrm{dDst}/\mathrm{d}t$ and $V$. The Ap and AE indices attain peaks about half a day before the maximum of $V$, with correlation coefficients ranging from $\mathrm{cc}\approx0.6$ to $\mathrm{cc}\approx0.7$, depending on the sample used. The best correlations of Ap and AE are found with $\mathit{VB}_{s}$ with a delay of 3 h, being characterised by $\mathrm{cc}\gtrsim 0.6$. The Dst derivative $\mathrm{dDst}/\mathrm{d}t$ is also correlated with $\mathit{VB}_{s}$, but the correlation is significantly weaker $\mathrm{cc}\approx 0.4$ – 0.5, with a delay of 0 – 3 h, depending on the employed sample. Such low values of correlation coefficients indicate that there are other significant effects that influence the relationship between the considered parameters. The correlation of all studied geomagnetic parameters with $B$ are characterised by considerably lower correlation coefficients, ranging from $\mathrm{cc}=0.3$ in the case of $\mathrm{dDst}/\mathrm{d}t$ up to $\mathrm{cc}=0.56$ in the case of Ap. It is also shown that peak values of geomagnetic indices depend on the duration of the CIR-related structures. The Dst is closely correlated with Ap and AE ($\mathrm{cc}=0.7$), Dst being delayed for about 3 h. On the other hand, $\mathrm{dDst}/\mathrm{d}t$ peaks simultaneously with Ap and AE, with correlation coefficients of 0.48 and 0.56, respectively. The highest correlation ($\mathrm{cc}=0.81$) is found for the relationship between Ap and AE.

Appendices

Appendix A: CIR–HSS List

In Table 2 a list of the CIR–HSS structures identified in the two analysed periods is presented. The event label is defined in the first column, which is followed by the date and time of the CIR onset at the three-hour time resolution data. In the third and fourth column the beginning and the end of the CIR–HSS structure are displayed in the form of DOY, whereas the fifth column gives the corresponding duration of the structure, expressed in days. The question mark denotes that a given CIR–HSS was interrupted by ICME(s). In the last four columns the peak values (based on the three-hour resolution) of geomagnetic activity parameters are displayed.

Table 2 CIRs–HSSs in the analysed periods.

Appendix B: The Effect of Time Resolution

To demonstrate how the applied time resolution affects the correlation results, we illustrate the dependence of linear least-squares fit parameters on the time resolution. In particular, the time resolutions of 1 h, 3 h, 6 h, and 12 h are applied to the relationships $\mathrm{Dst}(V)$, $\mathrm{AE}(V)$, $\mathrm{Dst}(\mathit{VB}_{s})$, and $\mathrm{AE}(\mathit{VB}_{s})$. The dependence of the correlation coefficients, cc, are summarised in Figure 7 and Table 3. Note that a similar analysis of the time-resolution effects, including the Ap index, was performed by Verbanac et al. (2013).

Figure 7

Effect of the applied time resolution on the correlation coefficients for the relationships (a) $\mathrm{Dst}(V)$ and $\mathrm{AE}(V)$ and (b) $\mathrm{Dst}(\mathit{VB}_{s})$ and $\mathrm{AE}(\mathit{VB}_{s})$.

Table 3 Dependence of the correlation coefficients for $\mathrm{Dst}(V)$, $\mathrm{AE}(V)$, $\mathrm{Dst}(\mathit{VB}_{s})$, and $\mathrm{AE}(\mathit{VB}_{s})$ on the applied time resolution.

In the case of the $\mathrm{Dst}(V)$ and $\mathrm{AE}(V)$ relationships, the differences between the linear least-squares fits performed on data based on different time resolutions are completely insignificant – the slopes and $y$-axis intercepts differ by less than 1% from those shown in Figures 2a and c, respectively. The only difference is in the range covered by the data, since both the geomagnetic data and $V$ attain lower values at lower time resolutions. However, the main difference is in the significant increase of the correlation coefficients with decreasing time resolutions (Figure 7a), primarily due to the corresponding data smoothing. This is especially pronounced in the case of the $\mathrm{AE}(V)$ relationship (red line in Figure 7a), since higher resolution AE data are very structured.

The situation is quite different for the relationships $\mathrm{Dst}(\mathit{VB}_{s})$ and $\mathrm{AE}(\mathit{VB}_{s})$. In addition to the change in the data range, here the slopes of the linear least-squares fits also change, systematically increasing with decreasing time resolutions. This is illustrated in Figure 8, where the 1 h time resolution correlations (red line) are compared with those for the 12 h resolution (black line). This is basically caused by lower values of $\mathit{VB}_{s}$, being smoothed at lower resolutions. The change in correlation coefficients (Figure 7b) is not so well defined as in the case of the $\mathrm{Dst}(V)$ and $\mathrm{AE}(V)$ relationships displayed in Figure 7a.

Figure 8

Comparison of correlations at a time resolution of 1 h (red pluses and red line) and 12 h (black crosses and black line), shown for (a) $\mathrm{Dst}(\mathit{VB}_{s})$ and (b) $\mathrm{AE}(\mathit{VB}_{s})$.