Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://www.iki.rssi.ru/people/gma49.pdf Äàòà èçìåíåíèÿ: Tue Feb 1 16:03:29 2011 Äàòà èíäåêñèðîâàíèÿ: Mon Oct 1 23:41:43 2012 Êîäèðîâêà: Ïîèñêîâûå ñëîâà: m 11

ISSN 0016 7932, Geomagnetism and Aeronomy, 2011, Vol. 51, No. 1, pp. 49­65. © Pleiades Publishing, Ltd., 2011. Original Russian Text © N.S. Nikolaeva, Yu.I. Yermolaev, I.G. Lodkina, 2011, published in Geomagnetizm i Aeronomiya, 2011, Vol. 51, No. 1, pp. 51­67.

Dependence of Geomagnetic Activity during Magnetic Storms on the Solar Wind Parameters for Different Types of Streams
N. S. Nikolaeva, Yu. I. Yermolaev, and I. G. Lodkina
Space Research Institute, Russian Academy of Sciences, Profsoyuznaya ul. 84/32, Moscow, 117997 Russia e mail: nnikolae@iki.rssi.ru
Received December 10, 2009; in final form, August 10, 2010

Abstract--The dependence of the maximal values of the |Dst| and AE geomagnetic indices observed during magnetic storms on the value of the interplanetary electric field (Ey) was studied based on the catalog of the large scale solar wind types created using the OMNI database for 1976­2000 [Yermolaev et al., 2009]. An analysis was performed for eight categories of magnetic storms caused by different types of solar wind streams: corotating interaction regions (CIR, 86 storms); magnetic clouds (MC, 43); Sheath before MCs (ShMC, 8); Ejecta (95); Sheath (ShE, 56); all ICME events (MC + Ejecta, 138); all compression regions Sheaths before MCs and Ejecta (ShMC + ShE, 64); and an indeterminate type of storm (IND, 75). It was shown that the |Dst | index value increases with increasing electric field Ey for all eight types of streams. When electric fields are strong (Ey > 11 mV m­1), the |Dst | index value becomes saturated within magnetic clouds MCs and possibly within all ICMEs (MC + Ejecta). The AE index value during magnetic storms is independent of the electric field value Ey for almost all streams except magnetic clouds MCs and possibly the compressed (Sheath) region before them (ShMC). The AE index linearly increases within MC at small values of the electric field (Ey < 11 mV m­1) and decrease when these fields are strong (Ey > 11 mV m­1). Since the dynamic pressure (Pd) and IMF fluctuations (B) correlate with the Ey value in all solar wind types, both geomagnetic indices (|Dst| and AE) do not show an additional dependence on Pd and IMF B. The nonlinear relationship between the intensities of the |Dst| and AE indices and the electric field Ey component, observed within MCs and possibly all ICMEs during strong electric fields Ey, agrees with modeling the magnetospheric­ionospheric current system of zone 1 under the conditions of the polar cap potential saturation. DOI: 10.1134/S0016793211010099

1. INTRODUCTION Numerous experiments have indicated that mag netic storms are mainly caused by southward IMF (see, e.g., [Russell et al., 1974; Burton et al., 1975; Akasofu et al., 1985; Gonzalez et al., 1999; O'Brien and McPherron, 2000; Vennerstroem, 2001; Yermo laev and Yermolaev, 2002; Sheath Lyatsky and Tan, 2003; Huttunen and Koskinen, 2004; Rusanov and Petrukovich, 2004; Maltsev, 2004; Veselovsky et al., 2004; Gonzalez and Echer, 2005; Yermolaev et al., 2005a, 2005b, 2007a] and references therein). The effectiveness of the IMF Bz component in the genera tion of magnetic storms and substorms depends on the interaction between the induced electric field Ey = VxBz (Vx is the radial component of the solar wind velocity at the southward IMF component Bz < 0) and the magnetospheric­ionospheric system. The distri bution of currents within this system changes as a result of these processes (see, e.g., [Gonzalez et al., 1994; O'Brien and McPherron, 2000; Siscoe et al., 2005; Zhu et al., 2006] and references therein), which affects the magnetic field value on the Earth's surface
49

and leads to changes in the value of the geomagnetic indices. The intensity of a magnetospheric disturbance is estimated using the Dst and AE indices. The high lati tude AE index characterizes the intensity of the auroral electrojet current and is the indicator of substorm activity [Davis and Sugiura, 1966]. The low latitude Dst index is used to estimate the ring current intensity during magnetic storms and is the measure of the interplanetary disturbance geoeffectiveness [Sugiura, 1964; Burton et al., 1975]. We emphasize that auroral activity during magnetic storms has been studied. Sub storms without magnetic storms possibly develop dif ferently: e.g., AlfvÈn waves can cause substorms but cannot cause storms (see [McPherron et al., 2009] and references therein). In addition, during magnetic storms auroral electrojets shift to lower latitudes rela tive to the latitude of the stations used to determine the AE index, and results can be distorted [Feldstein, 1992]. The electric field Ey is the main solar wind param eter for the generation of magnetic storms and sub storms. At the same time, the magnetic storm intensity


50

NIKOLAEVA et al.

is also affected by variations in other solar wind parameters: dynamic pressure Pd and IMF fluctu ations [Burton et al., 1975; Gonzalez et al., 2001, 2002; Borovsky and Funstein, 2003; Seki et al., 2005; Yermolaev et al., 2007a, 2007b]. The relative contribu tion of the dynamic pressure Pd to the Dst index value is not a constant value as followed from the work (see [Burton et al., 1975]). The effect of Pd on the Dst index value depends on the value of the electric field component Ey and the contribution of Pd is insignifi cant at the large Ey values usually observed during the main storm phases [Siscoe et al., 2002, 2005]. An ordinary quasistationary solar wind does not contain a considerable and prolonged IMF Bz compo nent sufficient for the generation of a magnetic storm. At the same time, certain disturbed types of solar wind streams can have a large and prolonged IMF Bz com ponent, including such a component oriented south ward, which results in a magnetic storm [Tsurutani et al., 1988, 1995; Yermolaev and Yermolaev, 2002; Huttunen and Koskinen, 2004; Huttunen et al., 2002; Richardson et al., 2002; Vieira et al., 2004; Echer and Gonzalez, 2004; Yermolaev et al., 2005a, 2010a]. Several geoeffective solar wind streams, during the passage of which magnetic storms with different inten sities are observed, can be distinguished. Interplane tary manifestations of coronal mass ejections CMEs (ICME events including two subtypes: magnetic clouds MCs and Ejecta) and the regions where high speed solar wind streams interact with slow streams (CIRs) are the main types of these streams. The stud ies indicated that the region of compressed plasma (Sheath) can be observed before the ICME forward front (before MC, ShMC or Ejecta "piston", ShE). The nature of the Sheath formation is close to that of CIRs; however, ICME plays the role of a "piston" in this case instead of a high speed solar wind stream. Each stream type has its own set of solar wind parameters different from other stream types. For example, at the front of high and low speed streams in (CIR events) and before the "piston" forward front (Sheath), plasma has increased values of density and tempera ture, and the thermal pressure predominates over the magnetic one > 1. A magnetic cloud is a subclass of ICME events and has a higher (>10 nT) and more reg ular magnetic field than Ejecta. Both subtypes of ICME events (MC + Ejecta) have a magnetic field structure in the form of a rope and the magnetic pres sure predominates over the thermal one within them ( 1). Note that we do not consider the class of very strong magnetic storms that are generated by several ICMEs interacting with one another [Yermolaev and Yermolaev, 2008]. The method for identifying differ ent types of solar wind streams based on the OMNI plasma and magnetic database for 1976­2000 is described in detail in [Yermolaev et al., 2009]. The complete statistics of the solar wind events are pre sented and their geoeffectiveness as the magnetic

storm occurrence probability (i.e., the ratio of the number of events that caused a magnetic storm to the total number of events of a given type) is estimated in [Yermolaev et al., 2010a]. It is known that the develop ment of geomagnetic storms and substorms substan tially differs depending on the stream type that caused these phenomena [Borovsky and Denton, 2006; Des pirak et al., 2009]. In particular, these differences are observed in the behavior of the ring current, aurora, and Earth's plasma sheet; in magnetospheric convec tion; and in the polar cap potential saturation [Borovsky and Denton, 2006]. The works devoted to the relationship between the Dst index minimum and the IMF Bz component mostly ignore the types of interplanetary disturbances that generated magnetic storms (see, e.g., [Akasofu et al., 1985; O'Brien and McPherron, 2000] and refer ences therein). The situation is similar concerning the dependence of the AE index on Ey (see, e.g., [Weimer et al., 1990]). Only some works have considered the relationship between these parameters in specific types of solar wind streams: MCs [Wu and Lepping, 2002; Yurchyshyn et al., 2004; Yermolaev et al., 2007b], corotating interaction regions CIR [Alves et al., 2006; Richardson et al., 2006; Yermolaev et al., 2007b], or streams behind interplanetary shocks [Oh and Yi, 2004]. The number of the works comparing these depen dences in different stream types with different internal structures is even smaller. The characteristics of the electric field Ey variations during the passage of two main solar wind types (MCs and CIRs) were com pared in [Kershendolts et al., 2007; Plotnikov and Barkova, 2007]. The authors indicated that a nonlin ear dependence of the AE and |Dst | indices on Ey is observed for MCs. These indices linearly increase with increasing electric field at small values (Ey < 12 mV m­1). When the field is large (Ey > 12 mV m­1), the |Dst | index becomes saturated, and the AE index decreases [Kershendolts et al., 2007; Plotnikov and Barkova, 2007]. At the same time, both indices linearly vary in the entire Ey range for CIR. The authors also studied the relationship of the indices with the dynamic pres sure and IMF fluctuation level observed in MC and CIR. In this case the maximal hourly values of the |Dst | and AE indices within MC and CIR were compared with the Pd value and IMF fluctuation level in Sheath before MC. The authors found that the relationships between the indices and solar wind parameters are sig nificant for CIR (r > 0.5) and are insignificant for MC (r < 0.5). The internal structure of MCs themselves and their possible division into two parts (the Sheath before MC and the MC body) was factually ignored in these works [Kershendolts et al., 2007; Plotnikov and Barkova, 2007] when the dependences of the indices on Ey were studied. In our previous work [Yermolaev et al., 2007b], we considered the region of ICME itself (i.e., the sum of
Vol. 51 No. 1 2011

GEOMAGNETISM AND AERONOMY


DEPENDENCE OF GEOMAGNETIC ACTIVITY DURING MAGNETIC STORMS

51

MC and Ejecta) and the Sheath before them (ShMC + ShE) as two different types of solar wind streams. This work [Yermolaev et al., 2007b] compared the depen dences of the Dst index minimum on the electric field Ey component for four types of the solar wind events: ICMEs (MC + Ejecta), corotating solar wind streams CIRs, Sheath (ShMC + ShE), and events of an indeter minate type. In addition, the achieved results were also compared with the data obtained by other authors [Wu and Lepping, 2002, 2005; Alves et al., 2006; Srivastava and Venkatakrishnan, 2004; Kane, 2005]. However, in [Yermolaev et al., 2007b] we used only linear depen dences for regression lines and did not numerically estimate the correlation ratio (correlation coeffi cients). The present work continues the previous one [Yer molaev et al., 2007b], but we decreased the Dst thresh old at a storm minimum from ­60 to ­50 nT in order to increase the magnetic storm statistics; i.e., all mod erate magnetic storms were analyzed. Second, we divided all moderate magnetic storms into eight groups (instead of four groups) depending on the type of a solar wind stream that caused a magnetic storm. The main aim of this work is to study the correlations between the magnetic storm intensities (the Dst index minimum and the |Dst | value are used) and substorm activity (the AE index) during a storm and the main solar wind parameters: the Ey component of the solar wind electric field, dynamic pressure Pd, and IMF fluctuation level . We analyze the values of the geo magnetic indices and solar wind parameters observed at a peak of the magnetic storm main phase. The main novelty of our paper consists in that we compare the response of the magnetosphere to different interplan etary drivers, the list of which is presented most com pletely at the modern stage of science. This approach should be first tested using the simplest method for comparing interplanetary and magnetospheric condi tions and then developed based on more physically perfect concepts (we have continued working in this field and published certain data on studying the dynamics of development of geomagnetic indices and solar wind parameters during the entire main phase of magnetic storms [Yermolaev et al., 2010b]). Our method for characterizing storms using the minimal negative value of the Dst index is one of the simplified methods used to compare magnetic storms and their geoeffectiveness with the solar wind. For example, it is better to characterize the complete magnetic storm intensity by comparing the total geomagnetic index for the storm period with the solar wind electric field dur ing the storm period or with the Akasofu parameter; however, this subject is outside the scope of this study. According to our method of analysis, it is implicitly assumed that the time of the magnetospheric response to the solar wind is not more than 1 h. At the same time, studying the dynamics of development of the geomagnetic indices and solar wind parameters during
GEOMAGNETISM AND AERONOMY Vol. 51 No. 1

the entire magnetic storm main phase indicates that a magnetic storm "remembers" previous Bz (and Ey) values [Yermolaev et al., 2010b]. Geomagnetic indices can theoretically correlate with many solar wind parameters; however, the number of independent parameters affecting indices can be small since these parameters are interrelated. The number of these parameters and what parameters should be used is the problem to be solved in the present study. 2. METHOD The list of magnetic storms with the minimum of Dst ­50 nT during the 1976­2000 period, for which the source was found in the solar wind [Yermolaev et al., 2009] was used as initial data for studying the interrelation between the Dst and AE geomagnetic indices, the solar wind parameters (the electric field Ey component and dynamic pressure Pd), and the IMF B fluctuation level. The method for identifying dif ferent types of the solar wind stream related to mag netic storms were described in detail in [Yermolaev et al., 2009, 2010a]. The list of magnetic storms with Dst 50 nT was completed with the data on the fol lowing quantities: the dynamic pressure Pd, IMF B fluctuation level, and AE index at a minimum of the magnetic storm Dst index (i.e., we used the AE index value at the instant when Dst was minimal averaged over 1 h rather than the maximal AE value). The AE index values averaged over 1 h were taken from the OMNI 2 database (http://omniweb.gsfc.nasa.gov/ form/dx1.html) [King and Papitashvili, 2004]. In addition, we estimated the Ey component of the con vective electric field (Ey = VxBz, where Vx is the radial component of the solar wind velocity at the negative IMF Bz component) for the instant when Dst was min imal. All moderate and strong magnetic storms with Dst ­50 nT registered in 1976­2000 were divided into eight groups depending on the stream type that caused these storms. As a result, it was found that 43 magnetic storms were related to MCs, 86 storms were related to CIR, 95 storms were related to Ejecta, 56 storms were related to the Sheath before Ejecta (ShE), 8 storms were related to the Sheath before MC (ShMC), 64 storms were related to all Sheath events before ICME (ShMC + ShE), 138 storms were related to all ICMEs (MC + Ejecta), and 75 magnetic storms were related to an indeterminate type of the solar wind stream (IND). The indeterminate type included the events for which it was impossible to reliably identify the stream type, or some parameters were absent, or the phenomenon type was complex. Note that we did not perform a linear approxima tion when the number of events was smaller than 10 because reliability for a small number of events (e.g., for ShMC when we analyzed the dependence of AE and Dst on Ey, Pd, and IMF ). When the number of
2011


52
IND, CIR,

NIKOLAEVA et al.
Ejecta, MC+Ejecta, r1 = 0.97 ShE, ShMC+ShE, MC, ShMC |Dst| = 10.2 E y + 38.7

150

()
|Dst|, nT

100

50

0
IND, CIR, Ejecta, AE = 44.2 E y + 538.3

5
MC+Ejecta, r1 = 0.96 ShE,

10
ShMC+ShE, MC, ShMC

(b) 1000
AE, nT

500

0

5

Ey, mV m­1

10

15

Fig. 1. The dependence of the (a) |Dst |, and (b) AE indices on the Ey field averaged for the solar wind types marked by dif ferent symbols: IND, CIR, Ejecta, MC + Ejecta, ShE, ShMC + ShE, MC, and ShMC. Vertical and horizontal bars crossing each point show the rms deviation. Solid lines show the regression lines. The regression equations and the linear correlation coeffi cients (r1) are presented at the top of each panel.

events was smaller than 30, we did not perform a qua dratic approximation (e.g., for MC events when we analyzed the dependence of AE on Ey). For a quadratic empirical dependence, when the deviation from the linear dependence is characterized by a small number of points (e.g., one point locating far from the area of points), we performed an approximation by a broken line from two linear segments: segment 1 included many points, and segment 2 included the last two points from segment 1 plus these isolated points. 3. ANALYSIS RESULTS Figure 1 presents the dependence of the geomag netic indices averaged for each solar wind type on the average electric field Ey: (a) |Dst | and (b) AE. Dif ferent types of the solar wind stream are shown by dif ferent symbols (the corresponding denotations are presented above each plot). Vertical and horizontal bars crossing points show the rms deviation. The linear correlation coefficient (r1) and regression equation (shown by a solid line) are presented at the top of each panel. Figure 1 indicates the average intensity of magnetic storms and substorms during storms caused by differ ent solar wind streams and the relationship of this

intensity to the average electric field value . We can assume that ShMC are the most geoeffective streams according to the average electric field value Ey since these streams mainly result in intense magnetic storms (|Dst | > 100 nT) and strong auroral currents (AE > 1000 nT) [Echer et al., 2006]. The same events (ShMC) have the largest electric field values (Ey > 10 mV m­1). Also strong magnetic storms with Dst < 100 nT but with moderate auroral currents (500 < AE < 1000 nT) are caused by Sheath before Ejecta (ShE) and ICME (ShMC + ShE) and by MCs themselves. These three types of events have the smaller (by a factor of 1.5­2) average values of the electric field (Ey = 6­8 mV m­1). Events (MC + Ejecta) are at the geoeffectiveness boundary between strong and moderate magnetic storms with moderate auroral currents (500 < AE < 1000 nT). Ejecta and especially CIR events are mostly related to moderate magnetic storms and moderate auroral currents. Both types of events have even smaller Ey values: > 4 and < 5 mV m­1, respectively. According to the values of the indices, IND events are less geoeffective and are mostly related to moderate storms and moderate auroral electrojet currents. The smallest average value of the electric field (Ey ~ 4mV m­1) corresponds to this type of events. Thus,
Vol. 51 No. 1 2011

GEOMAGNETISM AND AERONOMY


DEPENDENCE OF GEOMAGNETIC ACTIVITY DURING MAGNETIC STORMS 2000 N = 24 1500 1000 500 0 5 () 2000 N = 56 1500 1000 500 0 5 (c) r1 = 0.41

53

AE, nT

r2 = 0.6

|Dst|, nT

100 0 10 15 20 Ey, mV m­1 Pd 3.0 (N = 20); r2 = 0.88 Pd < 3.0 (N = 23); r2 = 0.88 MC 5

|Dst|, nT

10 15 20 Ey, mV m­1 Pd 3.0 (N = 12); r2 = 0.72 Pd < 3.0 (N = 12); r2 = 0.63 (b) N = 43 200 r2 = 0.88

AE, nT

10 15 20 Ey, mV m­1 Pd 4.9 (N = 29); r1 = 0.38 Pd < 4.9 (N = 27); r1 = 0.14 (d) N = 86 200 r1 = 0.65 100 0 10 15 20 Ey, mV m­1 Pd 4.7 (N = 44); r1 = 0.53 Pd < 4.7 (N = 42); r1 = 0.73 CIR 5

Fig. 2. The dependence of the (b), (d) |Dst | and (a), (c) AE geomagnetic indices on the electric field value Ey for (a), (b) MC and (c), (d) CIR events.

each solar wind type is possibly characterized by its own average value of the electric field Ey and |Dst | and AE indices. The dependence of the AE and |Dst | geomagnetic indices on the electric field value Ey is shown in Figs. 2­5 separately for each group of magnetic storms related to the specific type of the solar wind stream. Solid lines show an approximation of the dependence or regression lines with the correlation coefficient values (r1 and r2 are the linear and quadratic correlation coefficients, respectively). Figure 2 shows the dependence of the (a), (c) AE and (b), (d) |Dst | indices on Ey for (a), (b) MCs and (c), (d) CIR events. Figures 2a and 2b indicate that the AE and Dst indices nonlinearly depend on the electric field Ey for MCs (the quadratic correlation coefficients are high: r2 = 0.6 and 0.88, respectively). We can assume that the values of the AE and |Dst | indices lin early increase at small field values (Ey < 11 mV m­1). However, at large Ey values (>11 mV m­1), the AE index decreases and the |Dst | index remains unchanged, i.e., indices are saturated. For clearness, we present in Fig. 2a an approximation of the AE dependence on Ey by a broken line including two lin ear segments (thin lines): segment 1 corresponds to Ey < 11 mV m­1 (the regression line constructed based on 21 points AE = 67.8Ey + 429.3); segment 2, to Ey > 10.5 mV m­1 (the regression line for four points AE = ­50.9Ey + 1568.4). Figure 2b similarly shows an approximation of the |Dst | dependence by a broken line composed of two linear segments (thin lines): seg
GEOMAGNETISM AND AERONOMY Vol. 51 No. 1

ment 1 with Ey < 11 mV m­1 (the regression line con structed using 39 points |Dst | = 15.1Ey + 14.7); seg ment 2 with Ey > 10.5 mV m­1 (the regression line for five points |Dst | = 0.59Ey + 200.9). For CIR events, the AE index is almost independent of the electric field value (the linear correlation coefficient is low r1 = 0.41), whereas the |Dst | index linearly increases with increasing electric field Ey with a correlation coeffi cient of r1 = 0.65, which indicates that these parame ters are closely related to each other (Figs. 2c, 2d). Figure 3 shows the dependence of the AE (Figs. 3a, 3c) and |Dst | (Figs. 3b, 3d) indices on the Ey field val ues observed in Ejecta events (Figs. 3a, 3b) and in the Sheath before these events (ShE; Figs. 3c, 3d). Figure 3a indicates that the AE index is independent of the electric field value for Ejecta events (the correlation coefficient is low r2 = 0.45). At the same time, the |Dst | index linearly increases with increasing Ey, and the linear correlation coefficient is high r1 = 0.82 (see Fig. 3b). For clearness, in Fig. 3a we show the linear approximation in two electric field intervals by thin straight lines: interval 1 for Ey < 11 mV m­1 with the regression line constructed based on 68 points AE = 50.6Ey + 493.56; interval 2 corresponds to Ey > 9mV m­1 with the regression line constructed based on three points AE = ­130.1Ey + 2255.4. It is clear that the AE index tends to decrease with increasing Ey > 11 mV m­1 as for MC events; however, the statis tics of Ejecta events is too insufficient and the correla tion coefficient is low (r2 < 0.5). Figures 3c and 3d
2011


54 2000 N = 69 1500 1000 500 0 ()

NIKOLAEVA et al. 2000 N = 46 1500 1000 500 0 (c) r1 = 0.03

AE, nT

r2 = 0.45 10 15 m­1 35); r2 = 0.39 34); r2 = 0.55 r1 = 0.82

5 Ey, mV Pd 3.5 (N = Pd < 3.5 (N = N = 95 (b)

AE, nT

5 10 15 Ey, mV m­1 Pd 6.0 (N = 22); r1 = 0.0004 Pd < 6.0 (N = 24); r1 = 0.27 (d) r1 = 0.60 5 10 15 Ey, mV m­1 Pd 5.65 (N = 30); r1 = 0.59 Pd < 5.65 (N = 26); r1 = 0.63 ShE

|Dst|, nT

|Dst|, nT

200 100 0

300 N = 56 200 100 0

5 10 15 Ey, mV m­1 Pd 3.3 (N = 48); r2 = 0.79 Pd < 3.3 (N = 47); r2 = 0.84 Ejecta

Fig. 3. The dependence of the (b), (d) |Dst | and (a), (c) AE indices on the electric field value (Ey) for (a), (b) Ejecta and (c), (d) Sheath before Ejecta (ShE).

indicate that strong electric fields with Ey > 11 mV m­1 are more frequently registered in the Sheath before Ejecta (ShE) than in the Ejecta region (seven and one events, respectively). For ShE events, the AE index is independent of the Ey field value (the correlation coef ficient is low r1 = 0.03); at the same time, the |Dst | index linearly increases with increasing Ey field and the corre lation coefficient is high r1 = 0.6 (see Figs. 3c, 3d). Figures 4a and 4b show the dependence of the AE and |Dst | indices on the electric field Ey for Sheath before MC (ShMC). Although the linear correlation coefficients are high (r1 = 0.76 and 0.86, respectively), the statistical significance of the result is low, since such events are rare. Figures 4c and 4d show the same for Sheath before all ICMEs (ShMC + ShE). It is evi dent that the AE index is independent on the Ey field for the (ShMC + ShE) sum of events and ShE events (the correlation coefficient is low r1 = 0.33). At the same time, the |Dst | index linearly increases with increasing electric field for these events (the correlation coeffi cient is high r1 = 0.7). Figures 5a and 5b show the same for all ICME events, i.e., for MC + Ejecta. It is clear that the AE index is independent of the electric field (the quadratic correlation coefficient is r2 = 0.47). The AE index dependence is approximated by a broken line com posed of two linear segments: segment 1 corresponds to Ey < 11 mV m­1 with the regression line constructed for 89 points AE = 60.9Ey + 455.0; segment 2 corre sponds to Ey > 10 mV m­1 with the regression line con

structed for five points AE = ­57.5Ey + 1545.6 and indicates that the AE index tends to increase at Ey < 10­11 mV m­1 and subsequently decreases when Ey > 11 mV m­1 as in MC events. At the same time, the |Dst| index linearly increases with increasing Ey for the same MC + Ejecta events (the linear correlation coefficient is r1 = 0.78). For illustration, Fig. 5b also shows an approximation by a broken line composed of two linear segments (gray lines): segment 1 with Ey < 11 mV m­1 (the regression line constructed for 134 points |Dst | = 12.9Ey + 29.1); segment 2 with Ey > 10.5 mV m­1 (the regression line constructed based on five points |Dst | = 0.67Ey + 198.8). The statistics of the events with Ey > 11 mV m­1 are insignificant (eight MC + Ejecta events have the Ey field larger than 10 mV m­1); therefore, we can only state that the ring current intensity tends to saturate when the Ey electric fields are large. Figures 5c and 5d show the same for IND events. For these events, the AE index value is independent of the Ey field; at the same time, the |Dst | index linearly depends on the electric field with a high correlation coefficient (r1 = 0.70). To verify the possible effect of the variations in the solar wind dynamic pressure Pd and in the level of IMF fluctuations on the dependence of the indices on the Ey field, we divided each type of event into two subgroups depending on the Pd value (circles for a low dynamic pressure Pd < P0 and crosses for a high pres sure Pd P0, where P0 is the dynamic pressure thresh old value shown in the captions below Figs. 2­5) and on the IMF value (circles for a low fluctuation level
Vol. 51 No. 1 2011

GEOMAGNETISM AND AERONOMY


DEPENDENCE OF GEOMAGNETIC ACTIVITY DURING MAGNETIC STORMS 2000 N=5 1500 1000 500 0 5 () AE, nT 2000 N = 51 1500 1000 500 0 5 (c) r1 = 0.33

55

AE, nT

10 Ey, mV Pd 7 (N = Pd < 7 (N = N=8

15 20 25 m­1 3); r1 = 0.47 2); r1 = 1 |Dst|, nT

|Dst|, nT

300 200 100 0

(b)

10 15 20 25 Ey, mV m­1 Pd 6.1 (N = 26); r1 = 0.33 Pd < 6.1 (N = 25); r1 = 0.22 (d) N = 64 r1 = 0.70 300 200 100 0 10 15 20 25 Ey, mV m­1 Pd 6.1 (N = 31); r2 = 0.67 Pd < 6.1 (N = 33); r2 = 0.62 ShMC + ShE 5

10 15 20 25 Ey, mV m­1 Pd 7 (N = 4); r1 = 0.82 Pd < 7 (N = 4); r1 = 0.97 ShMC

5

Fig. 4. The dependence of the (b), (d) |Dst | and (a), (c) AE indices on the electric field value Ey for the following events: (a), (b) Sheath before MC (ShMC) and (c), (d) all Sheaths (ShE + ShMC).

< 0 and crosses for considerable fluctuations 0, where 0 is the threshold value of the IMF fluctuation level shown below Figs. 6­9). The P0 and 0 threshold values used to divide the interval were determined approximately so that the number of points in each subgroup would be identical or change insignificantly. For either subgroup with a high (crosses) and low (circles) pressure Pd, we determined the regression lines (not shown for shortness) and estimated the cor relation coefficients, the values of which are presented at the bottom of Figs. 2­5. An analysis of the sub groups with different dynamic pressure values gave the following results. First, a higher dynamic pressure Pd > P0 (compare circles and crosses in Figs. 2­5) is mostly observed at strong electric fields (Ey > 10 mV m­1) for all stream types (e.g., MC, ShMC, Ejecta, ShE, MC + Ejecta, ShMC + ShE, and IND) except CIR. At the same time, both low and high Pd values are observed at weaker fields (Ey < 10 mV m­1). Second, the correlation coefficient between the |Dst | index and field Ey is higher in the subgroup with low pressure (Pd < P0) than in the subgroup with high pressure (Pd > P0) or in the entire sample for some types of events (CIR, Ejecta, ShE, and ShMC). In contrast, the correlation is higher in the subgroup with high pres sure (Pd > P0) for other types of events (ShMC + ShE, MC + Ejecta, and IND). A similar analysis for each type of events was per formed for the subgroups with high ( 0) and low ( < 0) levels of IMF fluctuations. The
GEOMAGNETISM AND AERONOMY Vol. 51 No. 1

solid lines in Figs. 6­9 show the regression lines for all events (i.e., repeating Figs. 2­5), and the dotted lines show the regression lines for the subgroup with high ( 0, crosses) and low ( < 0, circles) levels of IMF fluctuations. It is clear that strong electric fields (Ey > 10 mV m­1) are accompanied by a higher level of fluctuations for the MC, ShMC, MC + Ejecta, and ShMC + ShE events. A comparison of the correla tion coefficients for the subgroups with different levels of IMF fluctuations indicates that considerable IMF fluctuations make the relationship between the AE index and field Ey slightly closer for some types of events (e.g., MC, ShE, ShMC + ShE, MC + Ejecta, and IND) and between the |Dst | index and Ey (for ShE, ShMC + ShE, and IND events). To quantitatively verify the assumptions that high pressure (Pd P0) and considerable IMF fluctuations ( 0) can affect the dependence of the geomag netic indices on the Ey field during magnetic storms, we estimated the deviations of the AE and Dst indices from the regression lines for the groups with high and low Pd and with high and low levels of IMF fluctu ations. The results of these calculations and the num ber of events in either subgroup are presented in Figs. 10a­10c and Figs. 10d­10f for the |Dst | and AE indices, respectively. Figures 10b and 10e indicate that the |Dst | and AE indices deviate from the regression line for the sub groups with considerable ( 0) and insignificant ( < 0) IMF fluctuations, respectively. Figures 10c and 10f show the same for the subgroups with high
2011


56 2000 N = 92 1500 1000 500 0 5 ()

NIKOLAEVA et al. 2000 N = 57 1500 1000 500 0 5 (c) r1 = 0.41

AE, nT

r2 = 0.47 10 15 20 Ey, mV m­1 Pd 3.3 (N = 47); r2 = 0.41 Pd < 3.3 (N = 45); r2 = 0.25 (b) r1 = 0.78

AE, nT

10 15 20 Ey, mV m­1 Pd 3.3 (N = 28); r2 = 0.51 Pd < 3.3 (N = 29); r2 = 0.10 (d) r1 = 0.70

300 N = 138 |Dst|, nT 200 100 0

|Dst|, nT 10 15 20 Ey, mV m­1 Pd 3.15 (N = 68); r1 = 0.81 Pd < 3.15 (N = 70); r1 = 0.69 MC + Ejecta 5

300 N = 75 200 100 0

10 15 20 Ey, mV m­1 Pd 3.15 (N = 39); r1 = 0.81 Pd < 3.15 (N = 36); r1 = 0.006 IND

5

Fig. 5. The dependence of the (b), (d) |Dst | and (a), (c) AE indices on the electric field value Ey for the following events: (a), (b) all ICMEs (MC + Ejecta) and (c), (d) solar wind of an indeterminate type (IND).

(Pd > P0) and low (Pd < P0) dynamic pressures, respectively. An analysis of Fig. 10 indicates that the average deviations of both indices from the main dependence in either subgroup are substantially smaller than the rms deviations for all types of streams except ShMC, the number of which is too small. 4. DISCUSSION OF RESULTS The correlation between the solar wind parameters and geomagnetic indices has been analyzed in many works [Snyder et al., 1963; Gonzalez et al., 1998; Badruddin, 1998; Wang et al., 2003; Kane, 2005] (see also [Yermolaev et al., 2007b] and references therein). However, only some works took account of the solar wind stream type that caused a storm. Our results confirm that the dusk dawn electric field Ey component is mainly responsible for the values of the |Dst | and AE indices during moderate and strong magnetic storms, i.e., being the main geoeffective fac tor. On average, the Ey field can differ for different solar wind stream types, which depends on the stream formation physics and criteria of selection of different stream types. The dependence of the |Dst | index on pressure in some solar wind types is related to the fact that pressure directly correlates with field Ey in these solar wind types. That is, the index may actually depend only on one independent parameter Ey and indirectly depends on pressure. All information on the dependences of the AE and |Dst | indices on the electric field Ey we obtained,

including the regression equations and linear (r1) and quadratic (r2) correlation coefficients, for eight types of geoeffective solar wind events is presented in Tables 1 and 2, respectively. The results (linear and quadratic approximations with correlation coefficients) obtained in [Kershendolts et al., 2007] are given in an individual raw (marked by *). Note that the authors of this work included all ICMEs and Sheath before them in the MC events. Nevertheless, it is clear that their approximations are similar to our dependences of |Dst | on Ey for the CIR and MC types of streams. An analysis of the correlation confirms that the value of the |Dst | index depends on the electric field value Ey during a magnetic storm for all eight types of solar wind streams: the correlation coefficient is sig nificant (>0.5) and varies from 0.6 for ShE events to 0.86 for MC and ShMC events. Since the number of ShMC events we analyzed is too small (five events, see Table 2), the reliability of the correlation coefficient between the storm intensity |Dst | and electric field Ey obtained for these events is low in spite of the fact that the correlation coefficient value is large (r1 = 0.86). For the remaining stream types, we can assume that the closest relationship between the |Dst | index and electric field Ey is observed for the ICME events: r1 = 0.86, 0.82, and 0.78 for MCs, Ejecta, and their sum (MC + Ejecta), respectively. The relationship between Ey and |Dst | is less close for all compressed regions, including the Sheath and CIR, and for IND events: r1 = 0.7 for all Sheaths and ShE + ShMC, r1 = 0.7 for IND events, r1 = 0.65 for CIRs, and r1 = 0.6 for ShE.
Vol. 51 No. 1 2011

GEOMAGNETISM AND AERONOMY


DEPENDENCE OF GEOMAGNETIC ACTIVITY DURING MAGNETIC STORMS 2000 N = 24 1500 1000 500 0 5 () AE, nT 2000 N = 56 1500 1000 500 0 5 (c) r1 = 0.41

57

AE, nT

r2 = 0.6

10 15 20 Ey, mV m­1 B 3.0 (N = 12); r2 = 0.76 B < 3.0 (N = 12); r2 = 0.57 N = 43 (b) |Dst|, nT

10 15 20 Ey, mV m­1 B 4.0 (N = 28); r1 = 0.29 B < 4.0 (N = 28); r1 = 0.56 N = 86 (d) r1 = 0.65

|Dst|, nT

200 100 0

200 100 0

r2 = 0.88

10 15 20 Ey, mV m­1 B 2.4 (N = 20); r2 = 0.86 B < 2.4 (N = 23); r2 = 0.91 MC

5

10 15 20 Ey, mV m­1 B 4.0 (N = 44); r1 = 0.36 B < 4.0 (N = 42); r1 = 0.78 CIR

5

Fig. 6. The same as in Fig. 2 but points are divided according to the IMF fluctuation.

Although the differences between the correlation coefficient values are insignificant, we can assume that the observed effect is related to a more regular behav ior of Bz (and Ey) in ICMEs since a magnetic storm "remembers" previous Bz (and Ey) values [Yermolaev et al., 2010b]. The values of the correlation coefficients we obtained agree with the published data for individual types of events. For example, Kane [2005] considered the relationship between the Dst index minimum and the electric field (VBs) for moderate and strong mag netic storms (59 magnetic storms for 1973­2003, the Dst scale from ­50 to ­500 nT). A correlation coeffi cient of 0.67 was obtained as a result of an analysis per formed independently on the solar wind stream type [Kane, 2005], which is close to the values we obtained for the Sheath and CIR compressed regions and inde terminate events. An analysis of 64 intense magnetic storms (Dst 85 nT) registered by the ACE satellite in 1997­2002, performed independently on the source type in the solar wind, indicated that the peaks of the electric field Ey correlate with the Dst index minimum (the correla tion coefficient is r = 0.87) [Gonzalez and Echer, 2005], which is close to our data obtained for MCs. Vivek, Gupta and Badruddin [2009] analyzed 88 intense magnetic storms with Dst < ­100 nT (to 370 nT) during cycle 23, caused by five different types of the solar wind streams. However, the depen dence of the Dst index on Ey is presented for all pow erful storms independently on the stream type that caused this storm. The relationship between Dst peaks
GEOMAGNETISM AND AERONOMY Vol. 51 No. 1

and Ey is linear (Dst = ­77.40Ey ­ 7.85) and is charac terized by a high correlation coefficient, which also almost coincides with the values we obtained for MCs. We should note that about 18 of 88 events had large electric field values (Ey > 10 mV m­1). However, Vivek Gupta and Badruddin [2009] considered only the lin ear relationship between Ey and the Dst index (rather than the quadratic or piecewise linear dependence for MCs in our work and in [Kershendolts et al., 2007]). This is possibly related to the fact that strong magnetic storms were caused by the "mixture" of streams (i.e., ICMEs interacting with one another [Yermolaev and Yermolaev, 2008]) rather than by their individual types. The relationship between the Dst index and Ey for CIR was studied in [Alves et al., 2006; Zhang et al., 2008]. It was indicated that the Dst index linearly depends on Ey (Dst = ­11.1Ey ­ 11.3 [Zhang et al., 2008]; i.e., the regression line is close to our line for CIR (see Table 2). The correlation coefficient obtained in [Alves et al., 2006] for CIR (r1 = 0.66) almost coincides with the value (r1 = 0.65) that we obtained for these events. Cane et al. [2000] also obtained a linear relationship between the maximal southward IMF and the Dst index with a correlation coefficient of r = 0.74 for the Ejecta events with the Sheath before them (Ejecta + Sheath), which is close to the correlation coefficient (0.78) that we obtained for all ICMEs (MC + Ejecta). Wu and Lepping [2002] studied the relationship between the magnetic storm intensity (the Dst minimum) and the electric field Ey component for 34 MCs during 1995­1998. The
2011


58 2000 N = 69 1500 1000 500 0 ()

NIKOLAEVA et al. 2000 N = 46 1500 1000 500 0 (c) r2 = 0.12

AE, nT

r2 = 0.45

|Dst|, nT

|Dst|, nT

200 100 0

5 10 15 Ey, mV m­1 B 2.4 (N = 35); r2 = 0.50 B < 2.4 (N = 33); r2 = 0.51 r1 = 0.82 (b) N = 95

AE, nT

5 10 15 Ey, mV m­1 B 4.5 (N = 22); r2 = 0.31 B < 4.5 (N = 24); r2 = 0.26 (d) r1 = 0.60 5 10 15 Ey, mV m­1 B 4.5 (N = 28); r1 = 0.63 B < 4.5 (N = 28); r1 = 0.55 ShE

300 N = 56 200 100 0

5 10 15 Ey, mV m­1 B 2.4 (N = 46); r1 = 0.81 B < 2.4 (N = 48); r1 = 0.83 Ejecta

Fig. 7. The same as in Fig. 3 but points are divided according to the IMF fluctuation.

2000 N = 5 1500 1000 500 0 5

() AE, nT r1 = 0.76

2000 N = 51 1500 1000 500 0 5

(c) r1 = 0.33

AE, nT

10 15 20 25 Ey, mV m­1 B 9.1 (N = 3); r1 = 0.87 B < 9.1 (N = 2); r1 = 1 N=8 (b) |Dst|, nT r1 = 0.86

10 15 20 25 Ey, mV m­1 B 4.5 (N = 27); r2 = 0.48 B < 4.5 (N = 24); r2 = ­0.24 Np = 64 (d) r1 = 0.70

|Dst|, nT

300 200 100 0

300 200 100 0

10 15 20 25 Ey, mV m­1 B 7 (N = 4); r1 = 0.66 B < 7 (N = 4); r1 = 0.78 ShMC

5

10 15 20 25 Ey, mV m­1 B 4.5 (N = 34); r1 = 0.74 B < 4.5 (N = 30); r1 = 0.55 ShMC + ShE

5

Fig. 8. The same as in Fig. 4 but points are divided according to the IMF fluctuation.

authors used only the linear relationship between the storm intensity and the electric field within MCs and obtained a correlation coefficient of r1 = 0.79, which is slightly smaller than the value that we obtained for MCs (r1 = 0.86) but almost coincides with the correla tion coefficient (r = 0.78) that we obtained for all

ICMEs (Ejecta + MC). Note that about five MCs had strong fields (Ey > 10 mV m­1); however, the relation ship between the parameters was described only by a linear dependence [Wu and Lepping, 2002]. A smaller value of the correlation coefficient for MCs than such a value that we obtained is possibly explained by this fact.
Vol. 51 No. 1 2011

GEOMAGNETISM AND AERONOMY


DEPENDENCE OF GEOMAGNETIC ACTIVITY DURING MAGNETIC STORMS 2000 N = 92 1500 1000 500 0 5 () AE, nT r2 = 0.47 2000 N = 57 1500 1000 500 0 5 (c) r1 = 0.41

59

AE, nT

|Dst|, nT

|Dst|, nt

10 15 20 Ey, mV m­1 B 2.4 (N = 43); r2 = 0.55 B < 2.4 (N = 48); r2 = 0.44 300 N = 138 (b) 200 r1 = 0.78 100 0 10 15 20 Ey, mV m­1 B 2.3 (N = 66); r1 = 0.70 B < 2.3 (N = 71); r1 = 0.86 MC + Ejecta 5

10 15 20 Ey, mV m­1 B 3.0 (N = 27); r1 = 0.46 B < 3.0 (N = 30); r1 = 0.30 (d) 300 N = 75 200 r = 0.70
1

100 0 10 15 20 Ey, mV m­1 B 3.0 (N = 37); r1 = 0.78 B < 3.0 (N = 38); r1 = 0.48 IND 5

Fig. 9. The same as in Fig. 5 but points are divided according to the IMF fluctuation.

The results of our analysis indicate that the ring current intensity |Dst | possibly nonlinearly depends on the electric field Ey component for MCs and maybe all ICMEs (MC + Ejecta): the quadratic correlation coefficients are high for these events. Specifically, the ring current linearly increases with increasing electric field when its values are small (Ey < 11 mV m­1); how ever, at large field values (Ey > 11 mV m­1), the satura tion of the |Dst | index is observed when the ring current intensity remains unchanged. A similar character of the relationship between the |Dst | index and Ey was obtained in [Kershendolts et al., 2007; Plotnikov and Barkova, 2007] for MCs with the Sheath before them (ShMC). Our results for MCs and possibly all ICMEs confirm the conclusion on the behavior of the indices for large electric fields (Ey > 12 mV m­1) drawn by these researchers. Tables 1 and 2 indicate that the AE and |Dst | indices are related to Ey by the quadratic dependence for MCs, which is close to the regression lines in [Kershendolts et al., 2007; Plotnikov and Barkova, 2007] and has close correlation coefficients for the AE r2 = 0.7 (0.6 in our work) and |Dst | r2 = 0.9 (0.88 in our work) indices. The CIR events are also characterized by a similar linear dependence of the Dst index on Ey with similar correlation coefficients r1 = 0.65 (0.7 in our work). Our results achieved for CIR events substantially differ from the data obtained in [Kershendolts et al., 2007; Plotnikov and Barkova, 2007]: r1 = 0.41 and 0.6, respectively; and the relation ship between the AE index and the Ey field is absent. Moreover, according to our data, the AE index is inde pendent of (or weakly depends on) the electric field Ey
GEOMAGNETISM AND AERONOMY Vol. 51 No. 1

component for almost all event types except rare ShMC events (i.e., including CIRs). The differences can be explained by different CIR selection criteria. An important result of our work consists determin ing that the dynamic pressure Pd and the level of IMF fluctuations do not affect the dependence of the |Dst | index during magnetic storms (the Dst minimum during the storm main phase) on the electric field Ey. We can assume that both geomagnetic indices do not depend on the dynamic pressure and IMF fluctuation level but possibly affect the scatter of points (the value of uncertainty). Additional studies should be per formed in order to make a final conclusion. This conclusion partially agrees with other works, which indicated that the contribution of Pd to the Dst index decreases with increasing geoeffective compo nent of the electric field (Ey) [O'Brien and McPher ron, 2000, 2002; McPherron and O'Brien, 2001]. Specifically, the coefficient before the dynamic pres sure b(E), which was used to calculate the corrected Dst* index in the formula presented in [Burton et al., 1975], is not constant. This coefficient decreases by a factor of five when the electric field Ey increases from 0 to 18 mV m­1 [O'Brien and McPherron, 2000, 2002; McPherron and O'Brien, 2001; Siscoe et al., 2005]. Consequently, the correction for Dst, related to an increase in the dynamic pressure during the magnetic storm main phases, is possibly smaller than is usually assumed in the formula [Burton et al., 1975; Siscoe et al., 2005], which indirectly agrees with the data we obtained. This is caused by a change in the current structure within the magnetosphere­ionosphere sys tem. MHD modeling indicates [Siscoe et al., 2002a;
2011


60 Number of points for|Dst| 60 40 20 0 50 0 -50

NIKOLAEVA et al. ()

Deviation for |Dst|, nT

(b) B < B0 B B0 (c)

Deviation for |Dst|, nT

50 0 -50 (d) 40 20 Pd < P0 Pd P0

Number of points for AE

0 800 400 0 -400 -800 400 0

Deviation for AE, nT

(e) B < B0 B B0

Deviation for AE, nT

(f) Pd < P0 Pd P0 MC CIR MC + Ejecta Ejecta ShE ShMC ShMC + IND ShE

-400

Fig. 10. The number of points and the deviation of the (a)­(c) |Dst | and (d)­(f) AE indices from the regression lines for different subgroups of events with Pd < P0 (thin line) and Pd P0 (thick line) (c), (f) and for the subgroups of events with < 0 (thin line) and 0 (thick line) (b), (e). Vertical bars show rms deviations. Panels (a) and (d) show the number of points in either subgroup for the |Dst | and AE indices, respectively.

White et al., 2001] that the Chapman­Ferraro (C­F) current system is replaced by the region 1 current sys tem when IMF has a strong southward component and a large electric field (Bz ~20 nT, Ey > 10 mV m­1). When the Ey electric fields are large, the region 1 cur rent increases with increasing solar wind pressure Pd [Siscoe et al., 2002b]. We factually have two modes of interaction between the solar wind and the magneto sphere: mode 1 (when the solar wind and C­F cur rents predominate) and mode 2 (when the ionosphere and region 1 currents predominate [Vasyliunas, 2004]). The dimensionless quantity (0pVA), which can be larger and smaller than unity, is responsible for this division. Here p is the ionospheric Pedersen con

ductance, VA is the AlfvÈn velocity, and is the effec tiveness of reconnection at the magnetopause, deter mined as a ratio of the reconnection potential to the maximum possible potential drop that can be imposed by the solar wind on the magnetosphere [Siscoe et al., 2005]. The magnetosphere demonstrates both interac tion types. Most of the time the interaction with the magnetosphere determined by the solar wind (i.e., dynamic pressure). However, the ionospheric mode predominates during the interaction with the mag netosphere (i.e., zone 1 currents predominate) during magnetic storms, when the AlfvÈn velocity (VA) is high and the effectiveness of reconnection at the magneto pause () is maximal. That is, when IMF has a strong southward component (Ey > 10 mV m­1), which takes
Vol. 51 No. 1 2011

GEOMAGNETISM AND AERONOMY


DEPENDENCE OF GEOMAGNETIC ACTIVITY DURING MAGNETIC STORMS

61

Table 1. Equations of regression between AE and Ey and the linear (r1) and quadratic (r2) correlation coefficients for differ ent types of solar wind SW type MC * CIR CIR* MC + Ejecta Ejecta ShE ShMC ShMC + ShE IND Number of points 24 35 56 38 92 69 46 5 51 57 AE AE AE AE AE AE AE AE AE AE AE AE AE AE AE AE AE = = = = = = = = = = = = = = = = = Regression equations 110.5 780.0 137.7 526.5 610.2 819.5 309.9 646.3 266.3 612.5 576.4 785.5 364.8 618.1 803.4 545.4 353.3 + + + + + + + + + + + + + + ­ + + 176.8Ey ­ 8.04Ey2 6.8Ey 186.9 ­ 7.4Ey2 53.7Ey 6.56Ey + 5.3Ey2 34.3Ey 126.8Ey ­ 6.3Ey2 18.3 155.4Ey ­ 10.28Ey2 22.3 66.2Ey ­ 3.96Ey2 3.87 60.99Ey 31.7Ey 13.87Ey + 2.39Ey2 58.4Ey 143.9Ey ­ 7.87Ey2 Correlation coef Correlation coef ficient r1 ficient r2 0.1 0.4 0.41 0.6 0.20 0.19 0.03 0.76 0.33 0.41 0.6 0.7 0.42 0.4 0.47 0.45 0.12

0.38 0.45

Note: Symbol * means that the data on this raw were taken from [Kershendolts et al., 2007].

Table 2. Equations of regression between |Dst | and Ey and correlation coefficients r1 and r2 for different types of solar wind SW type MC * CIR CIR* MC + Ejecta Ejecta ShE ShMC ShMC + Sh IND Number of points 43 35 86 38 138 95 56 8 64 75 |Dst | |Dst | |Dst | |Dst | |Dst | |Dst | |Dst | |Dst | |Dst | |Dst | |Dst | |Dst | |Dst| |Dst | |Dst | |Dst | |Dst | Regression equations = = = = = = = = = = = = = = = = = ­2.19 + 21.45Ey ­ 0.51Ey2 35.78 + 11.27Ey ­83.1 + 31.9Ey ­ 0.8Ey2 66.81 ­ 1.29Ey + 0.96Ey2 42.11 + 9.57Ey 31.8 + 9.8Ey 30.1 +13.4Ey ­ 0.12Ey2 36.39 + 11.32Ey 25.63 + 13.39Ey 32.01 + 10.92Ey +0.19Ey2 30.1 + 10.8 Ey 13.44 +19.04 ­ 0.84Ey2 48.21 + 9.28Ey 33.85 + 10.32Ey 36.96 + 9.51Ey ­ 0.03Ey2 37.0 + 9.92Ey 57.19 + 0.71Ey + 0.86Ey2 Correlation coeffi Correlation coeffi cient r2 cient r1 0.86 0.8 0.65 0.7 0.78 0.82 0.60 0.86 0.70 0.70 0.88 0.9 0.72 0.5 0.78 0.83 0.64

E

0.70 0.73

Note: Symbol * means that the data on this raw were taken from [Kershendolts et al., 2007].

place at the peak of the magnetic storm main phase (at the Dst index minimum), the C­F current system, which increases the Dst value (makes it less negative), is replaced by the region 1 current system, which is neutral with respect to the Dst index value [White et al., 2001; Siscoe et al., 2002a, 2005; Vasyliunas, 2004].
GEOMAGNETISM AND AERONOMY Vol. 51 No. 1

We emphasize that substorm activity AE during magnetic storms depending on the value of the solar wind electric field Ey was not studied previously for many different types of solar wind streams. The studies of eight different types of solar wind streams indicate that the auroral current intensity (the AE index)
2011


62

NIKOLAEVA et al.

depends on the Ey electric field value (the correlation coefficient is > 0.5) only for MCs and the Sheath before them (ShMC). For the remaining types of solar wind streams, the auroral current (the AE index) is independent on the Ey electric field value or this dependence is weak (the correlation coefficient is <0.5). The relationship between the AE index (auroral activity) and the IMF Bz component has long been known [Arnoldy, 1971; Meng et al., 1973; Baker et al., 1983; Weimer et al., 1990]. A statistical comparison of the AE index values with the IMF Bz component dur ing powerful geomagnetic storms (independently of the storm source type) indicated that the AE index tends to saturate (to a fixed AE value of ~1100 nT) when the electric fields are strong (Ey > 10 mV m­1) [Weimer et al., 1990]. According to our data for MC events, during magnetic storms the AE index first monotonously increases to a certain maximal value of 1000­1100 nT when the electric field is Ey < 11 mV m­1, which coincides with the data obtained in [Weimer et al., 1990]. However, when this index becomes satu rated, its value decreases at a subsequent increase in the field (Ey > 11 mV m­1), which differs from the result achieved in [Weimer et al., 1990]. This decrease is possibly related to the fact that the auroral electrojet moves equatorward during magnetic storms, and polar stations that are used to determine the AE index do not register it. The saturation of the AE index is caused by the fact that the polar cap potential becomes saturated because the magnetosphere­ionosphere coupling is the non linear process [Weimer et al., 1990]. The modeling of the magnetosphere­ionosphere coupling indicates that the level of the polar cap potential saturation decreases with increasing conductivity of the auroral ionosphere caused by diffuse precipitation [Weimer et al., 1990]. 5. CONCLUSIONS Using our Solar Wind Catalog created based on the OMNI database for 1976­2000, we analyzed the dependence of the Dst and AE geomagnetic indices (the intensities of the ring and auroral currents) on the value of the solar wind electric field Ey and dynamic pressure Pd and the level of IMF magnetic fluctu ations during the main phase of moderate and strong magnetic storms caused by different types of solar wind streams: CIR, MC, ShMC, Ejecta, ShE, MC + Ejecta, ShMC + ShE, and IND. We distinguished 363 moderate and strong mag netic storms with Dst ­50 nT, for which the follow ing sources were found in the solar wind: the com pressed region before high speed streams (CIRs, 86 magnetic storms); MCs (43 storms); Ejecta (95 storms); all ICME events (MC + Ejecta,

138 storms); Sheath before MCs (ShMC, 8 storms); Sheath before Ejecta (ShE, 56 storms); all Sheaths before MCs and Ejecta (ShMC + ShE, 64 storms); and events of an indeterminate type (IND, 75 storms). For each type of stream, we obtained the regression lines and estimated the correlation coefficients between the |Dst | and AE indices (the intensities of the ring and auroral currents) and the electric field (Ey). An analysis of the Dst index during the main phases of moderate and strong magnetic storms depending on the solar wind parameters indicated that (1) A linear relationship (with high correlation coefficients r1 > 0.5) between the |Dst | index and the electric field Ey is observed for all eight types of stream; (2) The |Dst | index becomes saturated (i.e., it reaches its maximal value) at large field values (Ey > 11 mV m­1) only for MCs (and possibly for the MC + Ejecta); and (3) Against a background of the Dst index depen dence on the Ey field, the value of the Dst index is apparently independent of the dynamic pressure Pd and the IMF fluctuation level. An analysis of the AE index during the main phases of moderate and strong magnetic storms depending on the solar wind parameters indicated the following: (1) The AE index is independent of (or weakly depends on) the electric field value (Ey) for almost all solar wind streams (the correlation coefficient is low, r < 0.5) except MCs and ShMC. (2) The AE index nonlinearly increases with increasing field Ey (the quadratic correlation coeffi cient is r2 = 0.6) for MC events and linearly increases (r1 = 0.76) for ShMC events (five events). For MCs, we can assume that the AE index linearly increases with increasing field Ey at small its values (Ey < 11 mV m­1) but possibly saturates and gradually decreases at strong electric fields (Ey > 11 mV m­1). Saturation of AE results from the limitation of the polar cap potential owing to the nonlinear nature of the ionosphere­ magnetosphere coupling [Weimer et al., 1990]. (3) Against the background of the dependence of the AE index on the Ey field, the AE index value is independent of (or weakly depends on) the dynamic pressure Pd and the IMF fluctuation level. ACKNOWLEDGMENTS We are pleased that we could use the OMNI data base. The OMNI data were obtained from the GSFC/SPDF OMNIWeb site http://omni web.gsfc.nasa.gov. This work was supported by the Russian Founda tion for Basic Research (project no. 07 02 00042) and by the Department of Physical Sciences, Russian Academy of Sciences (Program 15 Plasma Processes in the Solar System).
Vol. 51 No. 1 2011

GEOMAGNETISM AND AERONOMY


DEPENDENCE OF GEOMAGNETIC ACTIVITY DURING MAGNETIC STORMS

63

REFERENCES
Akasofu, S. I., Olmsted, N., Smith, E.J., Tsurutani, B., Okida, R., and Baker, D., Solar Wind Variations and Geomagnetic Storms: A Study of Individual Storms Based on High Time Resolution ISEE 3 Data, J. Geo phys. Res., 1985, vol. 90A, p. JA090, doi: 10.1029/JA090iA01p00325. Alves, M.V., Echer, E., and Gonzalez, W.D., Geoeffective ness of Corotating Interaction Regions as Measured by Dst Index, J. Geophys. Res., 2006, vol. 111, p. A07S05. Arnoldy, R.L., Signature in the Interplanetary Medium for Substorms, J. Geophys. Res., 1971, vol. 76, no. 22, pp. 5189­5201, doi: 10.1029/JA076i022p05189. Baker, D.N., Zwickle, R.D., Bame, S.J., Hones, E.W., Tsu rutani, B.T., Smith, E.J., and Akasofu, S. I., An ISEE 3 High Time Resolution Study of Interplanetary Parameter Correlations with Magnetospheric Activity, J. Geophys. Res., 1983, vol. 88A, pp. 6230­6242, doi: 10.1029/JA088iA08p06230. Borovsky, J.E. and Funsten, H.O., Role of Solar Wind Tur bulence in the Coupling of the Solar Wind to the Earth's Magnetosphere, J. Geophys. Res., 2003, vol. 108A, doi: 10.1029/2002JA009601. Borovsky, J.E. and Denton, M.H., Differences between CME Driven Storms and CIR Driven Storms, J. Geo phys. Res., 2006, vol. 111, p. A07S08, doi: 10.1029/2005JA011447. Burton, R.K., McPherron, R.L., and Russell, C.T., An Empirical Relationship between Interplanetary Condi tions and Dst, J. Geophys. Res., 1975, vol. 80, pp. 4204­ 4214. Cane, H.V., Richardson, I.G., and Cyr, O.C.St., Coronal Mass Ejections, Interplanetary Ejecta and Geomag netic Storms, Geophys. Res. Lett., 2000, vol. 27, no. 21, pp. 3591­3594. Davis, T.N. and Sugiura, M., Auroral Electrojet Activity Index AE and Its Universal Time Variations, J. Geophys. Res., 1966, vol. 71, pp. 785­803. Despirak, I.V., Lubchich, A.A., Yahnin, A.G., Kozelov, B.V., and Biernat, H.K., Development of Substorm Bulges during Different Solar Wind Structures, Ann. Geophys., 2009, vol. 27, pp. 1951­1960. Dmitriev, A.V., Crosby, N.B., and Chao, J. K., Interplane tary Sources of Space Weather Disturbances in 1997 to 2000, Space Weather, 2005, vol. 3, no. 3, p. 03001, doi:10.1029/2004SW000104. Echer, E. and Gonzalez, W.D., Geoeffectiveness of Inter planetary Shocks, Magnetic Clouds, Sector Boundary Crossings and Their Combined Occurrence, Geophys. Res. Lett., 2004, vol. 31, no. 9, p. L09808, doi: 10.1029/2003GL019199. Echer, E., Gonzalez, W.D., and Alves, M.V., On the Geo magnetic Effects of Solar Wind Interplanetary Mag netic Structures, Space Weather, 2006, vol. 4, p. S06001, doi: 10.1029/2005SW00200. Feldstein, Y.I., Modelling of the Magnetic Field of Mag netospheric Ring Current as a Function of Interplane tary Medium Parameters, Space Sci. Rev., 1992, vol. 59, pp. 83­165.
GEOMAGNETISM AND AERONOMY Vol. 51 No. 1

Gonzalez, W.D., Joselyn, J.A., Kamide, Y, Kroehl, H.W., and Rostoker, G., What is a Geomagnetic Storm?, J. Geophys. Res., 1995, vol. 99A, pp. 5771­5792. Gonzalez, W.D., Tsurutani, B.T., and de Gonzalez, A.L.C., Interplanetary Origin of Geomagnetic Storms, Space Sci. Rev., 1999, vol. 88, pp. 529­562. Gonzalez, W.D., Clua de Gonzalez, A., Sobral. J., et al., Solar and Interplanetary Causes of Very Intense Geo magnetic Storms, J. Atmos. Terr. Phys., 2001, vol. 63, pp. 403­412. Gonzalez, W.D., Tsurutani, B.T., Lepping, R.P., and Sch wenn, R., Interplanetary Phenomena Associated with Very Intense Geomagnetic Storms, J. Atmos. Terr. Phys., 2002, vol. 64, pp. 173­181. Gonzalez, W.D. and Echer, E., A Study on the Peak Dst and Peak Negative Bz Relationship during Intense Geo magnetic Storms, Geophys. Res. Lett., 2005, vol. 32, no. 18, p. L18103, doi: 10.1029/2005GL023486. Huttunen, K.E.J. and Koskinen, H.E.J., Importance of Post Shock Streams and Sheath Region as Drivers of Intense Magnetospheric Storms and High Latitude Activity, Ann. Geophys., 2004, vol. 22, pp. 1729­1738. Huttunen, K.E.J., Koskinen, H.E.J., and Schwenn, R., Variability of Magnetospheric Storms Driven by Differ ent Solar Wind Perturbations, J. Geophys. Res., 2002, vol. 107A, p. JA900171, doi: 10.1029/2001JA900171. Kane, R.P., How Good Is the Relationship of Solar and Interplanetary Plasma Parameters with Geomagnetic Storms?, J. Geophys. Res., 2005, vol. 110A, p. A02213, doi:10.1029/2004JA010799. Kershendolts, S.Z., Barkova, E.S., and Plotnikov, I.Ya., Dependence of Geomagnetic Disturbances on Extreme Values of the Solar Wind Ey Component, Geo magn. Aeron., 2007, vol. 47, no. 2, pp. 167­175 [Geo magn. Aeron. (Engl. Transl.), 2007, vol. 47, pp. 156­ 164]. King, J.H. and Papitashvili, N.E., Solar Wind Spatial Scales in and Comparisons of Hourly Wind and ACE Plasma and Magnetic Field Data, J. Geophys. Res., 2004, vol. 110A, p. A02209, doi: 10.1029/2004JA01804. Lyatsky, W. and Tan, A., Solar Wind Disturbances Respon sible for Geomagnetic Storms, J. Geophys. Res., 2003, vol. 108A, p. 2003, doi: 10.1029/2001JA005057. Maltsev, Y.P., Points of Controversy in the Study of Mag netic Storms, Space Sci. Rev., 2004, vol. 110, no. 3, pp. 227­277. McPherron, R.L. and O'Brien, T.P., Predicting Geomag netic Activity: The Dst Index, Geophys. Monogr. Am. Geophys. Union, 2001, vol. 155, pp. 339­345. McPherron, R.L., Kepko, L., and Pulkkinen, T.I., et al., Changes in the Response of the AL Index with Solar cycle and Epoch within a Corotating Interaction Region, Ann. Geophys., 2009, vol. 27, pp. 3165­3178. Meng, C.I., Tsurutani, D., Kawasaki, K., and Akasofu, S. I., Cross Correlation Analysis of the AE Index and the Interplanetary Magnetic Field Bz Component, J. Geo phys. Res., 1973, vol. 78, no. 4, pp. 617­629.
2011


64

NIKOLAEVA et al. 2006, vol. 111, p. A07S04, doi: 10.1029/2005HA011440. Vasyliunas, V.M., The Mechanical Advantage of the Mag netosphere: Solar Wind Related Forces in the Mag netosphere­Ionosphere­Earth System, Ann. Geo phys., 2007, vol. 25, pp. 255­269. Vieira, L.E.A., Gonzalez, W.D., Echer, E., and Tsurutani, B.T., Storm Intensity Criteria for Several Classes of the Driving Interplanetary Structures, Sol. Phys., 2004, vol. 223, no. 1 2, pp. 245­258, doi: 10.1007/s11207 004 1163 2. Vennerstroem, S., Interplanetary Sources of Magnetic Storms: A Statistical Study, J. Geophys. Res., 2001, vol. 106A, pp. 29175­29184. Veselovsky, I.S., Panasyuk, M.I., Avdyushin, S.I., et al., Solar and Heliospheric Phenomena in October­ November 2003: Causes and Effects, Kosm. Issled., 2004, vol. 42, no. 5, pp. 453­508 (Cos. Res., pp. 435­ 488). Vivek Gupta and Badruddin, Interplanetary Structures and Solar wind Behaviour during Major Geomagnetic Per turbations, J. Atmos. Terr. Phys., 2009, vol. 71, pp. 885­ 896. Weimer, D.R., Reinleitner, L.A., Kan, J.R., Zhu, L., and Akasofu, S. I., Saturation of the Auroral Electrojet Current and the Polar Cap Potential, J. Geophys. Res., 1990, vol. 95A, pp. 18981­18987. White, W.W., Schoendorf, J.A., Siebert, K.D., Maynard, N.C., Weimer, D.R., Wilson, G.L., Sonnerup, B.U.O., Siscoe, G.L., and Erickson, G.M., MHD Simulation of Magnetospheric Transport at the Mesoscale, Geo phys. Monogr. Am. Geophys. Union, 2001, vol. 125, pp. 229­240. Wu, C. C. and Lepping, R.P., Effect of Solar Wind Velocity on Magnetic Cloud Associated Magnetic Storm Inten sity, J. Geophys. Res., 2002, vol. 107A, p. 1346, doi: 10.1029/2002JA009396. Yermolaev, Yu.I. and Yermolaev, M.Yu., Statistical Rela tionships between Solar, Interplanetary, and Geomag netospheric Disturbances, 1976­2000, Kosm. Issled., 2002, vol. 40, no. 1, pp. 3­16 (Cos. Res., pp. 1­4). Yermolaev, Yu.I., Yermolaev, M.Yu., Zastenker, G.N., Zelenyi, L.M., Petrukovich, A.A., and Sauvaud, J. A., Statistical Studies of Geomagnetic Storm Dependen cies on Solar and Interplanetary Events: A Review, Planet. Space Sci., 2005a, vol. 53, no. 1 3, pp. 189­196. Yermolaev, Yu.I., Yermolaev, M.Yu., and Nikolaeva, N.S., Comparison of Interplanetary and Magnetospheric Conditions for CIR Induced and ICME Induced Magnetic Storms, EGU Geophys. Res. Abstr., 2005b, vol. 7, p. 01064. Yermolaev, Yu.I., Yermolaev, M.Yu., Lodkina, I.G., and Nikolaeva, N.S., Statistical Investigation of Helio spheric Conditions Resulting in Magnetic Storms, part 1, Kosm. Issled., 2007a, vol. 45, no. 1, pp. 3­11 (Cos. Res., pp. 1­8). Yermolaev, Yu.I., Yermolaev, M.Yu., Lodkina, I.G., and Nikolaeva, N.S., A Statistical Investigation of Helio spheric Conditions Resulting in Magnetic Storms, part 2, Kosm. Issled., 2007b, vol. 45, no. 6, pp. 489­498 (Cos. Res., pp. 461­470).
Vol. 51 No. 1 2011

O'Brien, T.P. and McPherron, R.L., An Empirical Phase Space Analysis of Ring Current Dynamics: Solar Wind Control of Injection and Decay, J. Geophys. Res., 2000, vol. 105A, pp. 7707­7719. O'Brien, T.P. and McPherron, R.L., Seasonal and Diurnal Variations of Dst Dynamics, J. Geophys. Res., 2002, vol. 107A, p. JA009435, doi: 10.1029/2002JA009435. Oh, S.Y. and Yi, Y., Relationships of the Solar Wind Param eters with the Magnetic Storm Magnitude and Their Association with the Interplanetary, J. Korean Astron. Soc., 2004, vol. 37, pp. 151­157. Plotnikov, I.Ya. and Barkova, E.S., Advances in Space Research Nonlinear Dependence of Dst and AE Indi ces on the Electric Field of Magnetic Clouds, Adv. Space Res., 2007, vol. 40, pp. 1858­1862. Richardson, I.G., Cane, H.V., and Cliver, E.W., Sources of Geomagnetic Activity during Nearly Three Solar Cycles (1972­2000), J. Geophys. Res., 2002, vol. 107A, p. JA000504, doi: 10.1029/2001JA000504. Richardson, I.G., Webb, D.F., Zhang, J., et al., Major Geo magnetic Storms (Dst < ­100 nT) Generated by Coro tating Interaction Regions, J. Geophys. Res., 2006, vol. 111, p. A07S09, doi: 10.1029/2005JA011476. Rusanov, A.A. and Petrukovich, A.A., Influence of Solar Wind Parameters on the Level of Geomagnetic Field Fluctuations, Kosm. Issled., 2004, vol. 42, no. 4, pp. 368­375 (Cos. Res. p. 354). Russell, C.T., McPherron, R.L., and Burton, R.K., On the Cause of Magnetic Storms, J. Geophys. Res., 1974, vol. 79, no. 7, pp. 1105­1109. Seki, T., Morioka, A., Miyoshi, Y.S., et al., Auroral Kilo metric Radiation and Magnetosphere­Ionosphere Coupling Process during Magnetic Storms, J. Geophys. Res., 2005, vol. 110, p. A05206, doi: 10.10292004JA010961. Siscoe, G.L., Crooker, N.U., and Siebert, K.D., Transpolar Potential Saturation: Roles of Region 1 Current System and Solar Wind Ram Pressure, J. Geophys. Res., 2002a, vol. 107A, p. JA009176., doi: 10.1019/2001JA009176. Siscoe, G.L., Erickson, G.M., Sonnerup, B.U.O., May nard, N.C., Schoendorf, J.A., Siebert, K.D., Weimer, D.R., White, W.W., and Wilson, G.R., Hill Model of Transpolar Potential Saturation: Comparisons with MHD Simulations, J. Geophys. Res., 2002b, vol. 107A, p. JA000109, doi: 10.1029/2001JA000109. Siscoe, G.L., McPherron, R.L., and Jordanova, V.K., Diminished Contribution of Ram Pressure to Dst dur ing Magnetic Storms, J. Geophys. Res., 2005, vol. 110, p. A12227, doi: 10.1029/2005JA011120. Srivastava, N. and Venkatakrishnan, P., Solar and Inter planetary Sources of Major Geomagnetic Storms dur ing 1996­2002, J. Geophys. Res., 2004, vol. 109A, p. A10103, doi: 10.1029/2003JA010175. Sugiura, M., Hourly Values of Equatorial Dst for the IGY, Ann. Int. Geophys. Year, 1964, vol. 35, pp. 945­948. Vasyliunas, V.M., Comparative Magnetospheres: Lessons for Earth, Adv. Space Res., 2004, vol. 33, doi: 10.1016/j.asr.2003.04.051. Vasyliunas, V.M., Reinterpreting the Burton­McPherron­ Russell Equation for Predicting Dst, J. Geophys. Res.,

GEOMAGNETISM AND AERONOMY


DEPENDENCE OF GEOMAGNETIC ACTIVITY DURING MAGNETIC STORMS Yermolaev, Yu.I. and Yermolaev, M.Yu., Comment on "Interplanetary Origin of Intense Geomagnetic Storms (Dst < ­100 nT) during Solar Cycle 23" by W.D. Gonzalez et al, Geophys. Res. Lett., 2008, vol. 35, p. L01101, doi: 10.1029/2007GL030281. Yermolaev, Yu.I., Nikolaeva, N.S., Lodkina, I.G., and Yer molaev, M.Yu., Catalog of Large Scale Solar Wind Phenomena during 1976­2000, Kosm. Issled., 2009, no. 2, pp. 99­113 (Cos. Res., pp. 81­94). Yermolaev, Yu.I., Lodkina, I.G., Nikolaeva, N.S., and Yer molaev, M.Yu., Statistical Study of Interplanetary Con dition Effect on Geomagnetic Storms, Kosm. Issled., 2010a, vol. 48, no. 6, pp. 461­474 (Cos. Res., pp. 485­ 500). Yermolaev, Yu.I., Nikolaeva, N.S., Lodkina, I.G., and Yer molaev, M.Yu., Relative Occurrence Rate and Geoef

65

fectiveness of Large Scale Types Solar Wind, Kosm. Issled., 2010b, vol. 48, no. 1, pp. 3­32 (Cos. Res., pp. 1­30). Yurchyshyn, V., Wang, H., and Abramenko, V., Correlation between Speeds of Coronal Mass Ejections and the Intensity of Geomagnetic Storms, Space Weather, 2004, vol. 2, p. S02001, doi: 10.1029/2003SW000020. Zhang, Y., Sun, W., Feng, X.S., Deehr, C.S., Fry, C.D., and Dryer, M., Statistical Analysis of Corotating Interac tion Regions and Their Geoeffectiveness during Solar Cycle 23, J. Geophys. Res., 2008, vol. 113, p. A08106, doi: 10.1029/2008JA013095. Zhu, D., Billings, S.A., Balikhin, M., Wing, S., and Coca, D., Data Derived Continuous Time Model for the Dst Dynamics, Geophys. Res. Lett., 2006, vol. 33, p. L04101, doi:10.1029/2005GL025022.

GEOMAGNETISM AND AERONOMY

Vol. 51

No. 1

2011