Positive ionospheric anomalies induced in the polar cap region by co-rotating interaction region (CIR)- and coronal mass ejection (CME)-driven geomagnetic storms are analysed using four-dimensional tomographic reconstructions of the ionospheric plasma density based on measurements of the total electron content along ray paths of GPS signals. The results of GPS tomography are compared with ground-based observations of F region plasma density by digital ionosondes located in the Canadian Arctic. It is demonstrated that CIR- and CME-driven storms can produce large-scale polar cap anomalies of similar morphology in the form of the tongue of ionization (TOI) that appears on the poleward edge of the mid-latitude dayside storm-enhanced densities in positive ionospheric storms. The CIR-driven event of 14–16 October 2002 was able to produce ionospheric anomalies (TOI) comparable to those produced by the CME-driven storms of greater Dst magnitude. From the comparison of tomographic reconstructions and ionosonde data with solar wind measurements, it appears that the formation of large-scale polar cap anomalies is controlled by the orientation of the interplanetary magnetic field (IMF) with the TOI forming during the periods of extended southward IMF under conditions of high solar wind velocity.
Geomagnetic storms result from substantial changes in solar wind pressure and the orientation of interplanetary magnetic field (IMF) caused by solar activity. During magnetic storms, enhanced coupling between the solar wind, IMF and the Earth’s magnetic field results in various magnetospheric phenomena such as intensification of the equatorial ring current (manifested by reductions in the Dst index), energization and precipitation of relativistic plasma particles, enhanced high-latitude plasma convection, etc. (e.g. Gonzalez et al. 1994). The largest geomagnetic storms characterized by Dst values below −100 nT (e.g. Zhang et al. 2007) are known to be generated by coronal mass ejections (CMEs), spontaneous eruptions of solar plasma causing abrupt enhancements in solar wind velocity accompanied by extended (few hours) periods of the southward IMF that facilitates the dayside magnetic reconnection and thus enhances the coupling between the solar wind and the magnetosphere.
Another mechanism known to produce geomagnetic storms that are distinct from those caused by CMEs involves the interaction between high- and slow-speed solar wind streams. High-speed solar wind streams are emitted from polar coronal holes and typically affect the Earth’s magnetosphere once every solar rotation period (27 days) causing recurrent co-rotating interaction region (CIR)-driven storms (e.g. Tsurutani et al. 1995). Statistically, the average solar wind speed of CME-related flows is found to be approximately 450 km s−1, which is an intermediate between the average of high-speed flows (approx. 500 km s−1) and average slow wind speeds of approximately 370 km s−1 (Richardson et al. 2002; Emery et al. 2009), although the solar wind speeds can be substantially higher in the initial hours of CMEs. Unlike the CME-driven disturbances, the CIR-driven storms do not produce a particularly strong ring current and show relatively mild reductions in the Dst index (with Dst typically remaining above −100 nT). Despite the weak Dst signatures, the CIR-driven storms can generate other magnetospheric phenomena that are typical of magnetic storms including the intensification of magnetosphere–ionosphere coupling (Tsurutani et al. 2006). While the duration of CME-driven storms typically does not exceed tens of hours, the CIR-driven storms can persist for a number of days (Denton et al. 2006). Consequently, the energy input into the magnetosphere during CIR-driven events can be comparable to, or even greater than, the energy input during CME-driven events (Turner et al. 2006). Owing to increased solar wind pressure, both types of storms are accompanied by long periods of enhanced magnetospheric convection. Comparing the effects of CME- and CIR-driven storms, Borovsky & Denton (2006) suggested that CIR-driven events are generally characterized by longer periods of the enhanced convection, even though the peak convection speed can be higher for CME-driven events.
The enhanced magnetospheric convection plays an important role in the re-distribution of plasma in mid-latitude, sub-auroral and high-latitude regions leading to various plasma anomalies generally known as positive ionospheric storms. Formation of the mid-latitude dayside ionospheric anomaly (also known as storm-enhanced density, or the SED anomaly) during large CME-driven storms has been extensively studied using combinations of the data of ionospheric GPS tomography, LEO spacecraft and incoherent scatter radars (Foster 1993; Foster et al. 2005; Mannucci et al. 2005). The ionospheric anomaly known as the tongue of ionization (TOI) spreads from the poleward edge of the mid-latitude SED anomaly across the auroral region into the polar cap. The TOI anomalies have been observed in the polar cap during CME-driven (Foster et al. 2005; Spencer & Mitchell 2007) as well as CIR-driven (Pokhotelov et al. 2009) events and they are believed to be linked to the plasmaspheric drainage plumes (Su et al. 2001; Foster et al. 2002) routinely observed in the magnetosphere during CME- and CIR-driven storms (Borovsky & Denton 2008). A statistical comparison between the TOI anomalies induced by CME- and CIR-driven events as well as the analysis of their connection to the drainage plumes has yet to be performed.
This paper will focus on the high-latitude ionospheric anomalies in the form of TOIs induced in the high-latitude ionosphere by typical CIR- and CME-driven storms. The CIR-driven storm that occurred during 14–16 October 2002 has been chosen as an example of the CIR-driven event. The ionospheric response to this CIR-driven storm has been described in detail in an earlier paper (Pokhotelov et al. 2009) and here we will focus primarily on the TOI anomaly appearing on the day of 16 October 2002. The geomagnetic storm of 6 November 2000 has been taken as an example of the CME-driven event. In order to reveal the ionospheric anomalies occurring in the polar cap and adjacent regions, the ionospheric tomography technique will be used which combines observations of total electron content (TEC) along multiple ray paths of GPS signals received by a network of ground-based GPS receivers to reconstruct three-dimensional distributions of plasma density. This tomographic technique has been used before to reconstruct ionospheric anomalies in the polar cap region during CME-driven (Spencer & Mitchell 2007; Pokhotelov et al. 2008) and CIR-driven (Pokhotelov et al. 2009) events, and an extensive verification of the technique has been performed using incoherent scatter radars (Yin et al. 2008) as well as digital ionosondes and in situ spacecraft data (Pokhotelov et al. 2010). The results of GPS tomography will be complemented by the data obtained by the Canadian Advanced Digital Ionosondes (CADI) located at the Eureka and the Resolute stations in the Arctic region of Canada.
2. Geomagnetic events analysed in the study
The geomagnetic storm of 14–16 October 2002 has been classified as being CIR-driven by Zhang et al. (2007). Solar wind parameters including the southward component of the IMF (Bz) and solar wind velocity (V sw) obtained by the ACE spacecraft, as well as the geomagnetic auroral electrojet (AE) and ring current intensity (Dst) indices for the period 14–16 October 2002, are presented in figure 1. Over this period, the solar wind velocity gradually increases from 300 to 600 km s−1 and the IMF Bz component shows oscillatory behaviour typical for CIR-driven storms (e.g. Tsurutani et al. 2006). As is common for CIR-driven events, this magnetic storm produces only a moderate equatorial signature with the Dst index reaching −100 nT around 14 UT on 14 October 2002 and two less pronounced Dst minima occurring on 15 October and 16 October. Ionospheric anomalies caused by the geomagnetic storm of 14–16 October 2002 at mid- and high latitudes have been previously analysed in detail in the context of global magnetosphere–ionosphere coupling (Pokhotelov et al. 2009). It has been shown that large-scale plasma anomalies in the form of TOIs appear in the polar cap on 14–16 October whenever the IMF Bz component remains negative for an extended period of time (e.g. few hours). This paper will only consider the ionospheric dynamics observed during 16 October when the magnitudes of high-latitude TEC anomalies were the largest. For convenience, solar wind parameters (IMF Bz, V sw) and geomagnetic indices (AE, Dst) during the 24 h of 16 October 2002 are shown in figure 2. It can be seen that the IMF remains southward during the interval 17–20 UT, which (under conditions of high solar wind speed) facilitates the formation of TOI spreading into the polar cap as will be demonstrated later.
The geomagnetic storm of 6 November 2000 has been classified (Zhang et al. 2007) as being caused by a CME. Solar wind parameters (IMF Bz, V sw) and geomagnetic indices (AE, Dst) during the 24 h of 6 November 2000 are presented in figure 3. The storm starts with the interplanetary shock arriving around 10 UT causing the solar wind velocity to jump from about 450 to 600 km s−1. The Dst index reaches its minimum value of −160 nT around 22 UT. The storm is characterized by an extended period of the southward IMF, which results in the formation of an unusually long-living TOI anomaly in the polar cap as will be demonstrated later by the GPS tomography and ionosonde measurements.
3. Ionospheric GPS tomography
Microwave signals from GPS navigational satellites experience group delay and phase advance proportional to the TEC along the signal path. Measurements of the TEC along multiple ray paths of GPS signals (so-called slant TEC) provide the input for the tomographic inversion algorithm in order to reconstruct the three-dimensional distribution of ionospheric plasma density (e.g. Bust & Mitchell 2008). In the high-latitude polar cap region where ground-based GPS receivers are sparsely distributed, the tomographic inversion algorithms may require some additional information about the plasma distribution. To fix the problem, this study uses the four-dimensional inversion algorithm known as MIDAS (Mitchell & Spencer 2003; Spencer & Mitchell 2007) in order to perform the tomographic inversion. The algorithm is based on a Kalman filter approach that accommodates temporal changes in the ionospheric plasma density by using the statistical Weimer model of global plasma convection (Weimer 1995) to project the ionospheric images forward in time. Thus, the use of a Kalman filter allows us to obtain three-dimensional density reconstructions over the areas where ground-based GPS receivers are sparsely distributed. Maps of two-dimensional vertical TEC distribution can be obtained by vertical integration of the plasma density through the tomographic three-dimensional reconstructions.
For this study, the tomographic reconstruction grid was centred over the north magnetic pole and covers a large portion of the Northern Hemisphere including the polar cap area, the auroral region and much of the mid-latitude region. The grid is uniform with 25×25×39 grid-cells in latitude, longitude and altitude, respectively (resolution of 4° in latitude, 4° in longitude and 40 km in altitude), covering the range of altitudes up to 1600 km. Some selected TEC reconstructions for the geomagnetic storms of October 2002 and November 2000 are shown in figures 4 and 5, respectively.
Figure 4 also shows the locations of 51 International GNSS Service (IGS) dual-frequency GPS receivers used for the tomographic reconstruction of the October 2002 event indicated by red labels with corresponding 4-symbol IGS station codes. The white x-cross indicates the location of the CADI ionosonde at the Eureka station (80° N, 274° E). The two panels in figure 4 show the reconstructed vertical TEC (1 TEC unit=1016 electrons m−2) at 18.00 UT (a) and at 19.30 UT (b). The first reconstruction shows the moment when the TOI starts to form at the northward edge of the mid-latitude TEC anomaly with plasma being transported anti-sunward by enhanced global convection flow. The second reconstruction shows the TOI fully developed and spreading into the polar cap region and over the north magnetic pole. Consequent TEC images (not shown here) demonstrate that the TOI finally separates from the mid-latitude plasma anomaly at around 20.30 UT and the remaining plasma drifts away across the polar cap into the nightside sector of the ionosphere. More detailed analyses of the TEC dynamics during this particular storm are presented in Pokhotelov et al. (2009). As mentioned earlier, the GPS tomography indicates that the TOI appears during this storm not only on 16 October but also on 14 and 15 October whenever the IMF Bz component remains negative for an extended period of time. The TEC magnitude of the TOI anomaly is largest on 16 October presumably owing to the higher solar wind speed, which facilitates faster polar cap anti-sunward convection.
The tomographic reconstructions during the CME-driven storm of November 2000 were based on the data from 40 dual-frequency GPS receivers located as shown in figure 5 by red labels with IGS station codes. The white x-cross indicates the location of the CADI ionosonde at the Resolute station (76° N, 265° E). The four panels in figure 5 demonstrate the vertical TEC maps at 14.30, 15.40, 16.40 and 20.30 UT. The colour scale is the same as in the reconstructions for the October 2002 storm presented earlier (0–80 TEC units). The time sequence of TEC images (not shown here) indicates that the TOI starts to form at the northward edge of the mid-latitude TEC anomaly around 14.00 UT. a (14.30 UT) shows the moment when the TOI is about to reach the Resolute station. b (15.40 UT) shows the TOI spreading over the north magnetic pole area into the nightside ionosphere following the anti-sunward convection flow. c (16.40 UT) shows how a large portion of the TOI separates from the mid-latitude anomaly and drifts away into the nightside sector while another portion of the TOI is still seen connected to the mid-latitude anomaly, spreading westward of the Resolute location. The TOI persists over a very long period during this storm and it can still be observed over the polar cap at 20.30 UT (figure 5d).
4. Digital ionosonde observations
In this study, data from the Canadian Advanced Digital Ionosonde (CADI) located at the Eureka station near the north magnetic pole is used to analyse the CIR-driven event of 16 October 2002. Unfortunately, the CADI dataset from Eureka for the CME-driven event of 6 November 2000 is incomplete and thus the data from another CADI ionosonde located at Resolute (about 500 km southward of Eureka) has been used for this event. CADI ionosondes produce conventional ionograms as well as provide velocity and ionospheric drift measurements derived from the fixed frequency Doppler measurements with temporal resolution of 30 s (MacDougall & Jayachandran 2001).
An example of CADI observations for the CIR-driven storm of 16 October 2002 from 16 to 22 UT is shown in figure 6. Figure 6a shows the reflection point of the ionospheric F region at a fixed frequency of 4.2 MHz. Figure 6b,c shows, respectively, the azimuth and the magnitude of the ionospheric convection flow. It can be seen that the anti-sunward ionospheric convection suddenly increases from 200 to about 1000 m s−1 around 17.40 UT, which is likely to be prompted by the sharp southward turning of the IMF (figure 2). This increase in convection is also confirmed by global observations of plasma convection with ionospheric SuperDARN HF radars (not shown here). From the analysis of TEC images presented earlier, it is clear that this particular moment corresponds to the time when the TOI starts to form at the northward edge of the mid-latitude TEC anomaly and to spread into the polar cap region. The dynamics of plasma transport over the Eureka station can be qualitatively understood by comparing the group range plot in figure 6 with the sequence of tomographic images shown in figure 4. After the ionospheric convection enhances at 17.40 UT and before the dense plasma from mid-latitudes reaches the north magnetic pole area at 18.50 UT, the F layer over the Eureka station remains relatively weak and inhomogeneous and we can see the passage of a few isolated ionospheric inhomogeneities as indicated by variations in the group range plot. After 18.50 UT, the ionospheric behaviour changes abruptly when the F layer becomes dense and uniform as indicated by constant reflection height values between 18.50 and 20.40 UT. This period corresponds to the passage of a dense TOI over Eureka as shown by the TEC map in figure 4a. After 20.40 UT, the F layer again becomes weak and inhomogeneous (as indicated by weak reflection and sharp variations in the reflection height) indicating that the TOI separated from the mid-latitude TEC anomaly and most of the enhanced plasma passed over the north magnetic pole drifting anti-sunward into the nightside ionosphere.
Since the ionosonde data provide an estimation of the bottom-side F layer density rather than the TEC value, it is preferable to compare the ionosonde measurements with the values of plasma density obtained from three-dimensional tomographic reconstructions. For this purpose, the vertical profile of plasma density over the Eureka station has been reconstructed from a time sequence of tomographic three-dimensional reconstructions. The resulting vertical density profile is shown as a function of time in figure 7a. As expected from the qualitative analysis above, we see an increase of the F layer density after plasma convection increases at 17.40 UT and the passage of dense TOI over Eureka is seen between 18.50 and 20.40 UT. The parameter most relevant to the ionosonde observations, the plasma density at the F layer peak, derived from the tomographic reconstructions is shown by the blue line in figure 7b, with the red line indicating the F peak density obtained from the relation ne=1.24×1010 (foF2)2 (e.g. Davies 1990) where the foF2 is the ordinary critical frequency (in MHz) measured by the CADI ionosonde. One can see that the GPS tomography replicates the enhancement of F peak density related to the passage of the TOI, though the density values derived from the tomography are somewhat (around 30%) lower than those measured by the ionosonde. That is typical for the tomographic algorithm used in this study as it tends to underestimate the F peak density while overestimating the density of the topside ionosphere (Pokhotelov et al. 2010).
CADI observations at the Resolute station during the CME-driven storm of 6 November 2000 from 13 UT to 23 UT are shown in figure 8. Figure 8a shows the reflection point of the ionospheric F region at the fixed frequency of 4.15 MHz. Figure 8b,c shows, respectively, the azimuth and the magnitude of the ionospheric convection flow. High values of the anti-sunward convection velocity (over 1000 m s−1) are seen around 14–15 UT when according to the GPS tomography the TOI forms at the northward edge of the mid-latitude anomaly and spreads towards the north magnetic pole. The moment when the TOI is about to reach the Resolute station (14.30 UT) is shown on the TEC map in figure 5a. From the CADI group range plot in figure 8 it can be seen that the ionospheric behaviour changes abruptly around 14.40 UT when the F layer becomes dense and uniform as manifested by constant reflection height values observed until about 16.20 UT indicating the passage of the TOI over Resolute. The period when the TOI is passing over the Resolute station and the north magnetic pole is shown on the TEC map in figure 5b (15.40 UT). The CADI group range plot in figure 8 shows that after 16.30 UT, the F layer again becomes weak and inhomogeneous indicating that the TOI separated from the mid-latitude TEC anomaly and the enhanced plasma passed over Resolute drifting anti-sunward into the nightside ionosphere. This is also shown in the TEC map in figure 5c (16.40 UT). However, it can be concluded from a sequence of TEC maps (not shown here) that the TOI does not disappear. Rather, it persists as a relatively continuous anomaly in the polar cap for another few hours as indicated by the TEC map in figure 5c (20.30 UT).
A more rigorous way of comparing the GPS tomography with CADI measurements is presented in figure 9. Here, the vertical profile of plasma density over the Resolute station has been reconstructed from a time sequence of tomographic three-dimensional images. The resulting vertical density profile is shown as a function of time in figure 9a. Figure 9b shows the F peak density derived from the tomographic reconstructions (blue line) and from CADI data (dotted red line). As concluded earlier from the comparison between TEC maps and group range CADI plots, the TOI arrives at Resolute around 14.40 UT and passes over the station until about 16.40 UT. After that, some enhanced plasma density is still seen over Resolute until the F layer becomes very weak around 18.30 UT. Then the F layer density again increases and the TOI re-appears (as indicated also by TEC maps in figure 5) and finally disappears around 22 UT.
It is interesting to compare the density reconstruction during the CME-driven event of 6 November 2000 (figure 9) with the behaviour of IMF Bz during this event shown in figure 3. The IMF turns sharply southward shortly after 13 UT and then the IMF Bz stays negative until 21.30 UT, except for the short interval when the IMF turns sharply northward around 18 UT. The behaviour of the F layer density seems to follow the IMF dynamics (e.g. high-density values under southward) with a delay of about 30 min, which may correspond to the convection time from the open-closed field line boundary. Figure 9 shows TOI at Resolute up to approximately 23 UT, which are remnants of TOI formed while the convective flow was strongly anti-sunward under the earlier Bz negative conditions. The same tendency can be observed when comparing the reconstructed density during the CIR-driven event of 16 October 2002 (figure 7) with the IMF behaviour during that day shown in figure 2. The IMF turns sharply southward after 17.20 UT and stays negative until 19.40 UT, while the enhanced plasma densities are seen over the Eureka station from 18 UT to 20.30 UT showing the delay time of 40–50 min, which is roughly the same as for the CME-driven event of 6 November 2000 (taking into account that Eureka is somewhat 500 km further north from Resolute). An estimation of the delay time can only be approximate as the convection velocity is unlikely to stay constant over many hours (as indicated by CADI convection measurements in figures 6 and 8).
5. Discussion and future work
Comparison between density anomalies induced by the CIR-driven storm of 16 October 2002 and the CME-driven storm of 6 November 2000 suggest that both events produce polar cap anomalies of similar magnitude and morphology. Sequences of TEC maps produced for both events reveal polar cap anomalies in the form of the TOI that persists for a number of hours spreading anti-sunward over the polar cap region. TEC magnitudes of the TOI anomalies are similar for the CIR- and the CME-driven storms with peak values around 80 TEC units. GPS tomography shows that the TOI associated with the CIR-driven event of October 2002 is relatively more uniform than the tongue associated with the CME-driven event of November 2000, while the TOIs exist for a substantially longer period during the CME-driven storm. However, we have no specific reason to believe that these are typical situations for all CME- and CIR-driven storms. Digital ionosonde measurements at Eureka (during October 2002 event) and Resolute (November 2000 event) as well as tomographic reconstructions of the vertical density profiles over these two sites indicate that both events produce plasma anomalies of similar magnitudes with maximum density values at the F peak being 1.7×1012 m−3 (1.2×1012 m−3) for the CIR-driven event of October 2002 and 2.1×1012 m−3 (1.1×1012 m−3) for the CIR-driven event of November 2000, as measured by CADI (reconstructed by the tomography). Taking into account that the Dst index reaches −160 nT during the CME-driven event of November 2000 and only goes down to −60 nT on 16 October 2002 (with a minimum Dst of −100 nT during the entire CIR-driven storm of 14–16 October 2002), it is reasonable to conclude that the CIR-driven storms can produce high-latitude ionospheric anomalies comparable to those produced by CME-driven storms of much greater Dst magnitude. This result implies that one cannot assess the intensity of storm-induced ionospheric anomalies like TOI by observing the Dst index alone. The interplanetary/solar origins of the geomagnetic storm must also be taken into account.
For both CIR- and CME-driven events, the dynamics of large-scale polar cap anomalies appears to be controlled by the orientation of IMF Bz component with the TOI forming only during extended periods of the southward IMF. This is confirmed by the fact that the TOI persists over the long (approx. 6 h) period of the southward IMF during the CME-driven storm of 6 November 2000, interrupted by the sharp northward IMF turning around 18 UT, though the moment of interruption is difficult to define owing to the fact that the magnitude of the TOI can be affected by variations in plasma convection velocity around 17–18 UT seen in the CADI data (figure 8). During the CIR-driven storm of 14–16 October 2002, the TOI appears only during the intervals when the IMF remained southward for over 2–3 h, e.g. before noon on 14 October and after noon on 15 and 16 October (Pokhotelov et al. 2009). During 16 October 2002, the interval analysed in the previous section, the TOI appears over Eureka after the IMF turns southward and disappears as the IMF turns northward with a delay roughly consistent with the convection time. The IMF control of polar cap plasma dynamics is not surprising since the southward IMF orientation facilitates the dayside reconnection thus creating a uniform plasma flow across the open-closed field line boundary into the polar cap. High solar wind velocity (driving faster polar cap convection) also appears to be favourable for the formation of the TOI, which is confirmed by the fact that largest TEC anomalies are seen in the polar cap during 16 October 2002 (e.g. when the solar wind velocity reaches its maximum) as opposed to 14 and 15 October 2002 (Pokhotelov et al. 2009). A less obvious conclusion comes from the fact that CIR-driven storms are characterized by oscillatory behaviour of the IMF Bz (Tsurutani et al. 2006) in contrast to CME-driven storms characterized by extended periods of southward IMF following the storm onset. Consequently, most CIR-driven storms are unlikely to create a long-lived TOI anomaly even though they may be characterized by very high solar wind velocities. This is of course a preliminary conclusion and a statistical study of many CIR-driven storms would be required to confirm it. Such statistical study is planned using the data from the CHAIN (Canadian High Arctic Ionospheric Network) network of digital ionosondes and GPS receivers recently deployed in the Canadian Arctic (Jayachandran et al. 2009).
While the question of the validity of the tomographic reconstructions in the polar cap region where GPS receivers are sparsely distributed is not addressed here, substantial work has been done on verifying the results of GPS tomography using CADI ionosondes, in situ spacecraft measurements (Pokhotelov et al. 2010), as well as incoherent scatter radars (Spencer & Mitchell 2007; Yin et al. 2008). In particular, the polar cap anomalies observed during the CIR-driven event of 14–16 October 2002 storm (especially during 16 October) have been validated using a combination of cold plasma measurements on-board DMSP (Defence Meteorological Satellite Programme) spacecraft and CADI ionosonde measurements (Pokhotelov et al. 2010). It has been demonstrated that the tomographic reconstructions successfully reproduce the large-scale (few hundred kilometres) spatial features of plasma anomalies detected by ionosonde and spacecraft measurements, although the GPS tomographic technique tends to underestimate the values of F layer peak density by 30–40%, while overestimating the density of the topside ionosphere at the 840 km DMSP orbit. This limitation of the GPS tomography is also evident when comparing tomographic reconstructions with CADI measurements presented here for both the CIR- and CME-driven events (figures 7 and 9). Care must be taken in the future in order to control the validity of tomographic reconstructions particularly when studying less extreme events producing anomalies of smaller magnitudes.
Ionospheric anomalies appearing in the high-latitude polar cap and adjacent regions during CIR- and CME-driven geomagnetic storms have been analysed using the four-dimensional GPS tomography technique and digital ionosonde observations in the polar cap. The geomagnetic storm of 14–16 October 2002 has been used as an example of a CIR-driven event while the storm of 6 November 2000 represents the case of a CME-driven event. In summary:
— It has been demonstrated that both CIR- and CME-driven storms can produce large-scale polar cap anomalies in the form of TOI spreading anti-sunward from the dayside mid-latitude plasma anomaly over the north magnetic pole into the nightside ionosphere following the global plasma convection pattern.
— Comparison between the sequences of tomographic TEC reconstructions for 16 October 2002 and 6 November 2000 suggest that the selected examples of CIR- and CME-driven storms produce anomalies of similar morphology and magnitude, though the TOI persists longer during the CME-driven event of 6 November 2000. The same applies to the comparison between the vertical density profiles reconstructed from the GPS tomography and digital ionosonde measurements of the F layer density at the Eureka and Resolute stations. It indicates that CIR-driven storms under certain favourable conditions can produce high-latitude ionospheric anomalies comparable to those produced by CME-driven storms of much greater Dst magnitude.
— The dynamics of polar cap anomalies induced by both the CIR- and CME-driven events appears to be controlled by the orientation of the IMF Bz component. The TOI forms during extended (few hours) periods of the southward IMF, when persistent convection carries plasma through the cusp into the polar cap, and disappears once the IMF turns northward. It suggests that most CIR-driven events may not produce a persistent TOI owing to the oscillatory behaviour of the IMF during such events.
— Statistical study of CIR-driven events is needed to clarify the relations between IMF/solar wind parameters and the dynamics/morphology of high-latitude ionospheric anomalies. Such a study will be performed using the network of digital ionosondes and GPS receivers recently deployed in the Canadian Arctic (CHAIN).
The Canadian authors would like to acknowledge the financial support from the Natural Sciences and Engineering Research Council of Canada. The UK authors acknowledge the financial support from the UK Science and Technology Facilities Council and the Royal Society Wolfson Research Award. CADI and CHAIN operations are conducted in collaboration with the Canadian Space Agency. We also would like to thank the International GNSS Service (IGS) for making the GPS data available and the NASA OMNIweb service for providing the solar wind data and geomagnetic indices.
One contribution of 8 to a Special feature ‘Geospace effects of high-speed solar wind streams’.
- Received February 9, 2010.
- Accepted April 7, 2010.
- © 2010 The Royal Society