SEARCH

SEARCH BY CITATION

Keywords:

  • magnetopause;
  • ring current

Abstract

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Data
  5. 3. LT Distribution of the GMCs
  6. 4. Dependence on the Solar Wind Conditions
  7. 5. Applications for a MP Model
  8. 6. Discussion
  9. 7. Conclusions
  10. Acknowledgments
  11. References

[1] Geosynchronous magnetopause crossing (GMC) data were collected from literature sources from 1967 to 1993 (189 GMCs) and from the experimental data on magnetic measurements on GOES (129 GMCs) and plasma measurements on LANL (197 GMCs) geosynchronous satellites in 1994 to 2001. The dawn-dusk asymmetry of the magnetopause at geosynchronous orbit was examined by two independent methods using the collected data set of 515 GMCs. First, the large amount of accumulated GMCs permitted the revealing of a substantial dawn-dusk asymmetry in the local time (LT) distribution of the GMC occurrence probability, with a statistically significant maximum in the range from 1000 LT to 1100 LT. Second, an analysis of the dawn-dusk asymmetry dependence on the upstream solar wind conditions was performed using a scatter plot of the solar wind total pressure versus local time for various IMF Bz. There was no asymmetry revealed for large positive Bz. Under strong negative Bz we found a prominent magnetopause dawn-dusk asymmetry. The asymmetry is characterized by a shifting of the GMCs with the minimal required solar wind total pressure toward the dawn and by a significantly lower (about 3 times) solar wind pressures required for the GMCs in the dawn sector relative to the dusk sector. We found that the asymmetry cannot be attributed to the IMF orientation along the Parker spiral, which is not revealed for strongly disturbed solar wind conditions accompanying the GMCs. An application of the dawn-dusk asymmetry effect for the Chao et al. [2002] model provided a substantial increase in the model predictive capability interim of the geosynchronous magnetopause crossings. The standard deviation decreased by 20% from 0.55 RE for the initial version to 0.45 RE for the asymmetrical version of the model, with the magnetopause axis rotated by an angle of about 15° toward the dawn. The physical processes responsible for the magnetopause dawn-dusk asymmetry are discussed. We indicate the two most probable magnetospheric phenomena, which would contribute to the substantial dawn-dusk asymmetry of the magnetopause under disturbed solar wind and geomagnetic conditions. The first one is that magnetopause erosion would operate more intensively in the prenoon sector. The second phenomenon is an asymmetrical terrestrial ring current that would develop during geomagnetic storms.

1. Introduction

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Data
  5. 3. LT Distribution of the GMCs
  6. 4. Dependence on the Solar Wind Conditions
  7. 5. Applications for a MP Model
  8. 6. Discussion
  9. 7. Conclusions
  10. Acknowledgments
  11. References

[2] Previous studies have provided contradictory information about the dawn-dusk asymmetry of the geosynchronous magnetopause crossings (GMCs). Some studies [Wrenn et al., 1981; Rufenach et al., 1989; McComas et al., 1993; Itoh and Araki, 1996; Kuznetsov and Suvorova, 1997, 1998a] report that GMCs are mostly observed on the dawnside, and therefore the magnetopause (MP) shape has dawn-dusk asymmetry during extreme solar wind (SW) conditions. However, McComas et al. [1994] present arguments against the asymmetry and show that the relatively small shift, of about half an hour in local time toward the morning for most GMCs, may be simply explained by the SW aberration effect. Table 1 represents a summary of previous studies of the GMCs. The first column indicates the paper where the GMC data set is presented. The second column shows the common number of the GMCs studied. Brief information about the experimental methods used for identification of the magnetopause crossings and the subject of the statistical investigation are presented in the third and fourth columns, respectively. Some authors perform a statistical analysis of the GMCs. That is, they analyzed only the magnetosheath (MS) entrances. Other statistical studies are based on the occurrence probability of the MS intervals, i.e., time intervals when the geosynchronous satellite is located in the magnetosheath. The last column of Table 1 presents an estimation of the MP asymmetry obtained in a study.

Table 1. Comparison of Studies of the GMC Dawn-Dusk Asymmetry
ReferenceGMC NumberGMC IdentificationSubject of AnalysisMP Asymmetry
  • a

    Determination of the exact number of GMCs is difficult (see the text).

  • b

    The data set contains 59 original GMCs and the leaving part is combination of the previous data sets.

Wrenn et al. [1981]15plasma dataMS intervals1000 LT
Rufenach et al. [1989]64magnetic dataGMCsprenoon
McComas et al. [1993]∼12aplasma dataMS intervalsprenoon
McComas et al. [1994]∼59aplasma dataMS intervals1130 LT
Itoh and Araki [1996]105magnetic dataGMCsprenoon
Kuznetsov and Suvorova [1997]172(59)bmixtureGMCs1100 LT

[3] Early data sets [Russell, 1976; Wrenn et al., 1981] contained only a few GMCs. Therefore they could not definitely reveal an asymmetry of the GMC occurrence. Wrenn et al. [1981] identified 15 magnetosheath intervals using the GEOS plasma observations. They found that the median of the local time distribution for the MS intervals is located at about 1000 LT. Obviously this result, based on very small statistics, has a very low confidence level.

[4] The first substantial data set, of 64 GMCs, was collected by Rufenach et al. [1989] from magnetic field observations on the GOES 2, 5, and 6 satellites in 1978 to 1986. The authors distinguished three types of GMCs in accordance to the SW conditions: (1) associated mostly with SW pressure enhancements (33 events), (2) caused by erosion of the geomagnetic field in the presence of southward Bz (6 events), (3) complicated events (25 events). Estimation of the dawn-dusk asymmetry was based on the statistical analysis of the local time distribution of the GMC occurrence probability, which was considered separately for each type of the GMCs. The slight skewing to prenoon hours was found for all GMCs and especially for the crossings associated with erosion and complicated events. The authors indicated the effect of asymmetry for only 31 GMCs that was again insufficient to make a certain conclusion.

[5] McComas et al. [1993] identified several GMCs using plasma measurements on the LANL geosynchronous satellites in 1992. On Plate 3 from the original paper we can distinguish 12 GMCs. The authors presented two independent instances of the dawn-dusk asymmetry. The LT distribution of the magnetosheath intervals has a prominent maximum in the prenoon sector at about 0900 LT. Simultaneous measurements of the GMCs by two LANL satellites separated in LT provide a case where the LANL moving in the prenoon sector detects the GMC while the LANL moving in the postnoon sector does not encounter the magnetosheath. The next study of the LANL observations of the GMCs [McComas et al., 1994] was based on 39 intervals selected in 1990 to 1993. Unfortunately, the GMCs from these intervals were not tabulated. We identified 59 GMCs using Figure 1 from the original paper. McComas et al. [1994] proposed three different methods for the examination of the magnetopause dawn-dusk asymmetry. The first one consisted of an analysis of the occurrence number of 5-min intervals when the LANL was located in the magnetosheath. A histogram of the MS interval occurrence numbers in 15-min bins of the LT was obtained. The distribution had a median value at 1130 LT with a standard deviation of 1 hour 22 min. The second method was multipoint simultaneous observations of the GMCs. In only one of the four cases the dawn-dusk asymmetry was revealed and estimated as an aberration of the magnetopause toward the dawn with a nose point located at ≤1145 LT. The third method addressed the location of the magnetopause nose, using the dawn-dusk component of the magnetosheath flow velocity, which should change sign just near the MP nose point. It was found that the median and mean local time for the nine flow reversals considered in the study were located very close to noon, with a standard deviation of about half an hour. Hence McComas et al. [1994] concluded that such a small dawn-dusk asymmetry of the magnetosheath intervals, magnetopause crossings and flow reversals, could be simply understood by taking into account the effect of SW wind aberration. However, the McComas et al. [1994] data set is still insufficient with regard to conclusion as a strict proof of MP symmetry at the geosynchronous orbit.

[6] Itoh and Araki [1996] collected 105 GMC events using magnetic data from GOES 2, 3, 5, 6, 7 during the time period from 1978 to 1992. The relation between the Dst index and the local time of the GMCs was studied. They found that in 16 events with positive Dst, the GMCs occurred only near local noon. Most of the GMCs (89 events) were observed under negative Dst variation and were located mostly in the prenoon sector.

[7] Kuznetsov and Suvorova [1997] used magnetic field measurements on GOES 7 to identify 59 GMCs from 1989 to 1993. The statistical analysis was also based on previous GMC data sets collected by Rufenach et al. [1989] and McComas et al. [1994]. Hence the analysis included in total 172 crossings. It was found that the local time distribution of the GMC occurrence probability has a strong dawn-dusk asymmetry with a median at 1100 LT. The asymmetry cannot be explained only by the SW aberration effect. Indeed, if we assume that the solar wind velocity has about the average value, 400 km/s, then the aberration angle is about 4°, which corresponds to about a 15 min shift of the MP nose toward the dawn. Actually, the GMCs are observed under strongly disturbed solar wind conditions when the solar wind velocity is much larger than its average value. Therefore for a simple SW aberration, the LT distribution of the GMC occurrence probability should have its median at 1145 LT or even closer to noon. On the other hand, the distribution obtained by Kuznetsov and Suvorova [1997] has a very wide profile and thus a very large statistical error. Hence their result can be considered only as an important indication on the GMC dawn-dusk asymmetry.

[8] The dawn-dusk MP asymmetry of the magnetopause can be represented by two modern magnetopause models [Kuznetsov and Suvorova, 1998b; Dmitriev and Suvorova, 2000]. The models were developed using the same data set of 39 GMCs. According to the study by Kuznetsov and Suvorova [1997], the magnetopause dawn-dusk asymmetry is very weak under moderately disturbed SW conditions but it has been found to be significant when the IMF Bz is negative and large (Bz < −6 nT) and/or the SW dynamic pressure is high (Pd > 20 nPa). The model by Kuznetsov and Suvorova [1998b] introduces the dawn-dusk asymmetry of the magnetopause by shifting the X-axis toward the dusk by ∼2 Earth radii (RE) for extreme SW conditions. The artificial neural network model by Dmitriev and Suvorova [2000] predicts a very large dawn-dusk magnetopause asymmetry at geosynchronous orbit, with the X-axis rotated at an angle θ ∼ 30° toward the dawn, that is equivalent to 0900 LT for the maximum of the GMC distribution.

[9] The main problem in the examination of the dawn-dusk magnetopause asymmetry at the geosynchronous orbit is insufficient GMC statistics. From Table 1 one can see that the largest GMC data set contained only 172 crossings [Kuznetsov and Suvorova, 1997]. In the present paper we introduce a new data set of 326 GMCs selected from GOES and LANL geosynchronous satellites in 1994–2001. The GMC selection method is described in section 2. Combining previous and new GMC data sets, we examine the dawn-dusk asymmetry of the GMCs by two independent methods. The first one is a statistical distribution of the GMC local time. It is presented in section 3. The second way, described in section 4, is the local time dependence of the minimal SW pressure required for the GMC. Application of the dawn-dusk asymmetry for the MP model is demonstrated in section 5. The results are discussed in section 6 and section 7 gives the conclusions.

2. Data

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Data
  5. 3. LT Distribution of the GMCs
  6. 4. Dependence on the Solar Wind Conditions
  7. 5. Applications for a MP Model
  8. 6. Discussion
  9. 7. Conclusions
  10. Acknowledgments
  11. References

[10] We analyze geosynchronous crossings of the dayside magnetopause (from 0600 LT to 1800 LT) in the time period from 1967 to 2001. We define a GMC as the magnetosheath encounter of the geosynchronous satellite. We also consider the MS intervals as when a geosynchronous satellite is located in the magnetosheath. From literature sources [Russell, 1976; Rufenach et al., 1989; McComas et al., 1994; Kuznetsov and Suvorova, 1997] 189 GMCs and 186 MS intervals were collected on the dayside from 1967 to 1993 (available at http://dec1.npi.msu.su/~alla/mp/gmc.html). To eliminate the SW aberration effect, the LT for each GMC is increased by 17 min, which is equivalent to an aberration angle of about 4°. In the present study we do not consider the solar wind plasma and IMF conditions for the GMCs from literature sources.

[11] Geosynchronous magnetopause crossings in the time period from 1994 to 2001 are selected. Magnetic field measurements on GOES 8, 9, 10 and plasma measurements on LANL (1990-095, 1991-080, 1994-084, LANL-97A, 1989-046) geosynchronous satellites are used. The method of GMC selection is described in detail by A. V. Suvorova et al. (Necessary conditions for the geosynchronous magnetopause crossings, submitted to Journal of Geophysical Research, 2004, hereinafter referred to as Suvorova et al., submitted manuscript, 2004). Here we present only a short description. High-resolution (∼1 min) ISTP data (available at http://cdaweb.gsfc.nasa.gov/cdaweb/istp_public/) of the geosynchronous satellites GOES and LANL and upstream monitors Geotail, Wind, and ACE are used. For a 1-min time resolution we suggest a practically instantaneous (1–2 min) response of the magnetopause, to changes of the interplanetary conditions influencing the magnetopause. A GMC is identified using the GOES magnetometer data, when one of two requirements is satisfied: (1) the magnetic field measured by the GOES deviates significantly from the geomagnetic field, (2) the magnetic field components measured by the GOES are correlated with the IMF components measured by an upstream monitor. Identification of the GMCs from the LANL data is based on a substantial increase of the low energy ion and electron content proper to the magnetosheath conditions. The GMC location is ordered in a fully aberrated GSM (aGSM) coordinate system where the X-axis is anti-parallel to the solar wind velocity.

[12] Valid determination of the upstream solar wind conditions corresponding to the GMCs, i.e., timing, is based on a time delay for direct propagation of the entire solar wind structure observed by an upstream monitor and on an additional time shift, which can be associated with tilted interplanetary fronts or with the evolution of the solar wind propagating from an upstream monitor to the magnetopause. The final timing is determined using two independent criteria. The first general criterion is based on a correlation between the SW pressure and the Dst (SYM-H) index that originates from solar wind pressure-associated changes of the Chapman-Ferraro current in the magnetopause [e.g., Burton et al., 1975; Russell et al., 1994a, 1994b]. The best correlation indicates the best timing as well as the best choice of the upstream solar wind monitor. The second, subsidiary, criterion for the upstream solar wind timing for the GOES satellites is co-variation of the Bz, By, and Bx components of the magnetic field measured by GOES during the magnetosheath intervals and the IMF components measured by an upstream solar wind monitor. For the LANL satellites, the independent criterion is the covariation of the ion density in the magnetosheath with the solar wind pressure. In such a way we select only GMC events, for which the magnetopause dynamics is directly associated with variations of the solar wind conditions or with changing of the geosynchronous satellite location. The accuracy of such methods, based on a ∼1-min time resolution of the experimental data, is estimated as a few minutes. The accuracy can be affected by a noninstantaneous magnetopause response, continuous changes of the solar wind front tilt [Collier et al., 1998], rapid variations of the solar wind plasma and IMF properties, and the evolution of the SW irregularities propagating through the interplanetary medium and the magnetosheath [Richardson and Paularena, 2001; Weimer et al., 2002].

[13] Using the method described above, we have identified 129 GMCs for the GOES satellites and 197 GMCs for the LANL satellites. The magnetosheath intervals observed by the GOES and LANL satellites contain 3004 and 2851 measurements, respectively. Hence the numbers of magnetosheath measurements by the GOES and LANL satellites from 1994 to 2001 are comparable, and thus our GMC data set can be considered as balanced, in the sense of experimental method for the GMC identification. Moreover, using different upstream monitors, we eliminate systematical errors in the upstream SW conditions associated with the GMCs. The total number of the GMCs and magnetosheath intervals obtained from the literature and selected from the experiment is 515 and 506, respectively.

3. LT Distribution of the GMCs

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Data
  5. 3. LT Distribution of the GMCs
  6. 4. Dependence on the Solar Wind Conditions
  7. 5. Applications for a MP Model
  8. 6. Discussion
  9. 7. Conclusions
  10. Acknowledgments
  11. References

[14] Figure 1 (left panel) shows the duration of the magnetosheath intervals (dT) versus local time. The closed and open symbols correspond to MS entrances and exits, respectively. The GMCs collected from the literature sources, the GOES and the LANL measurements are indicated by triangles, circles, and squares, respectively. The MS interval duration varies in a wide range from 1 min to ∼4 hours and GMCs cover completely the entire range of dT and LT variations. Practically in any region of the scatter plot of dT versus LT, one can find MP crossings belonging to each GMC data set (literature, GOES, and LANL). Hence the three different GMC data sets can be considered as being similar to one another, and thus we can combine them in one general set for statistical investigation. A histogram of the statistical distribution of the GMC-associated magnetosheath interval duration dT (right panel of Figure 1) is constructed using the combined data set of 506 MS intervals. The distribution has a maximum at dT ∼ 6 min. The median of the distribution (dotted line) is equal to 10 min and it falls closely to the bins with maximal statistics. The number of the GMCs decreases quickly with the duration dT, such that more than 70% of the MS intervals have a duration of less than 30 min and only one sixth of them have a duration of more than 1 hour.

image

Figure 1. (left) Distribution of the magnetosheath interval duration versus local time for the GMCs collected from the literature (triangles), GOES (circles), and LANL (squares) measurements. The closed and open symbols correspond to the magnetosheath entrances and exits, respectively. (right) Occurrence number of the MS interval duration. The median (10 min) is indicated by dotted line.

Download figure to PowerPoint

[15] The latitudinal distribution of the GMCs in the GSM coordinate system is presented in Figure 2 for the prenoon and postnoon sectors, respectively, indicated by solid and dashed histograms. Note that the exact location of the GMCs selected from the literature sources is not available. For these GMCs we assume the geographic latitude to be equal to Lat = 0°. There is no substantial difference in the shape of the GMC latitudinal distributions in GSM: their median (−5°) and maximum (−10°) are very close. However, the number of the GMCs observed in the prenoon sector (309) is much larger than the number in the postnoon sector (206). The difference indicates a dawn-dusk asymmetry in the GMC occurrence probability. As one can see from Figure 2, there is a similar distribution of the GMCs about the GSM equator on both sides of local noon. Hence the asymmetry is not associated with a geometrical effect of the asymmetry of the preferred latitudes for prenoon and postnoon sectors.

image

Figure 2. Statistical distribution of GMC latitudes in the GSM coordinate system for the prenoon (solid histogram) and postnoon (dashed histogram) sectors. Medians of the prenoon (−5°) and postnoon (−6°) distributions are indicated by the dotted and dashed lines, respectively. There is no substantial difference in the shape of the GMC latitudinal distributions. However, the number of GMCs observed in the prenoon sector (309) is much larger than the number of GMCs in the postnoon sector (206).

Download figure to PowerPoint

[16] A statistical distribution of the GMC occurrence number versus local time is presented in Figure 3. The width of the bins is equal to 1 hour. To compare different GMC data sets, we show the LT distributions separately for the GMCs selected from the literature (solid gray histogram), from the LANL, and GOES measurements (dashed gray and dashed black histograms, respectively). As we mentioned above, the GMC data sets are very similar to one another. In Figure 3 one can see that the statistical distribution for each data set has a maximum in the prenoon sector. The summarized histogram is presented by a solid line with error bars. There is a significat asymmetry of the LT distribution with the median (vertical dashed dotted line) and mean (vertical dashed line) equal to about 1120 LT. The bin containing maximal statistics N = 80 ± 9 occupies the range of LT from 1000 LT to 1100 LT. The statistics in this bin are higher by at least one standard deviation than the statistics in the bin at 1200 ∼ 1300 LT. The distribution can be well fitted by a Gaussian function with the center Xo at 1112 LT (vertical gray dotted line). Therefore we can conclude that the LT distribution of the GMC occurrence number has a statistically significant asymmetry with dawnward skewing of 1 hour, which is equivalent to an angle of 15°. The asymmetry cannot be explained by the SW aberration because this effect has already been taken into account in the GMC data sets.

image

Figure 3. Occurrence number of GMCs versus local time for the literature data (gray solid histogram), LANL crossings (gray dashed histogram), GOES crossings (black dashed histogram) and summarized (black solid histogram). The median (dashed dotted line), mean (dotted line) and center of the Gaussian approximation Xo (gray dotted line) are located in the prenoon sector at ∼1120 LT. The distribution has a statistically significant asymmetry with skewing toward dawn on ∼1 hour.

Download figure to PowerPoint

[17] Figure 4 shows the statistical distributions of the GMC local time separately for short magnetosheath intervals, with a duration of dT < 10 min (dashed histogram), and for long-lasting magnetosheath intervals, with dT ≥ 10 min (solid histogram). Note, that dT = 10 corresponds to the median and it is very close to the most probable duration of the MS intervals (see Figure 1). As one can see in Figure 4, both the distributions are asymmetrical relative to noon. Their medians are equal to 1142 LT and 1112 LT, respectively. The histogram for long-lasting MS intervals is shifted toward dawn more than the histogram for the short MS intervals. This difference could be attributed to the large width of the histogram bins. Indeed, because geosynchronous satellites rotate eastward, the GMCs at smaller local times could be preferred for long-lasting magnetosheath intervals. However, the magnetosheath intervals with a duration >1 hour contribute only to 30% of the statistics for long-lasting MS intervals and thus cannot significantly influence the skewing of the statistical distribution. Therefore the substantial dawn-dusk asymmetry of the magnetopause crossings at geosynchronous orbit is associated with neither the geometrical effects of the GMC selection nor the method of the GMC identification.

image

Figure 4. Statistical distributions of the GMC local time for short MS intervals with the duration dT < 10 min (dashed histogram) and for long-lasting MS intervals with dT ≥ 10 min (solid histogram). Their medians are indicated by the dashed (1142 LT) and dotted (1112 LT) vertical lines, respectively. Both the histograms are asymmetrical relative to noon.

Download figure to PowerPoint

4. Dependence on the Solar Wind Conditions

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Data
  5. 3. LT Distribution of the GMCs
  6. 4. Dependence on the Solar Wind Conditions
  7. 5. Applications for a MP Model
  8. 6. Discussion
  9. 7. Conclusions
  10. Acknowledgments
  11. References

[18] In this section we use only high time resolution data on the SW plasma and IMF conditions, which correspond to GMCs and magnetosheath intervals observed by the LANL and GOES satellites from 1994 to 2001, because in the literature only 1-hour averaged SW data are available. As we have shown in the previous section, the statistical properties of the literature, GOES, and LANL data sets are similar, so excepting the literature data set from consideration does not have very much affect on the generalizing of our results. For the time interval from 1994 to 2001, the geosynchronous satellites GOES and LANL provided data on 326 GMCs and 5855 MS measurements with a ∼1 min time resolution. Such great statistics allow for a detailed study of the solar wind conditions associated with the geosynchronous magnetopause crossings.

[19] Figure 5 displays the statistical distributions of the aberration angle in the GSM equatorial plane (about the Z-axis) δY (Figure 5a), the aberration angle in the GSM noon meridian δZ (about the aberrated Y-axis) (Figure 5b), the total SW pressure Psw (Figure 5c), and the IMF Bz (Figure 5d), which accompany the GMCs (solid histograms, left axis) and magnetosheath intervals (dashed histograms, right axis). A definition of the aberration angles δY and δZ was presented by Dmitriev et al. [2003]. The vertical dotted and dashed lines indicate the median of the distributions for the GMCs and MS intervals, respectively. The main statistical characteristics of the distributions are presented in Table 2. The dispersion is obtained by fitting the distribution with a Gaussian function. From Table 2, one can see that the distributions are very similar for the GMCs and MS intervals, despite a significant difference in the statistics. The median of the aberration angles δY and δZ is very close to 0°, which indicates that the solar wind propagation is practically radial in most GMCs. A small aberration for δY (about −3°) is associated with the Earth orbital rotation. Note that variations of the aberration angles are very large: the amplitude is up to 20° and the dispersion is about 4°. Hence accounting for the nonradial solar wind propagation and introducing the fully aberrated aGSM coordinate system is very important.

image

Figure 5. Statistical distributions of the solar wind conditions accompanying the GMCs (solid histograms, left axis) and magnetosheath intervals (dashed histograms, right axis): (a) aberration angle δY, (b) aberration angle δZ, (c) SW total pressure Psw, and (d) IMF Bz. Vertical dotted and dashed lines are medians of the distributions for the GMCs and MS intervals, respectively. The dotted histograms indicate common statistical distributions accumulated for (c) solar wind pressure (1-min time resolution) and (d) IMF Bz (16-s time resolution) measured by ACE in 1998 to 2001.

Download figure to PowerPoint

Table 2. Main Statistical Properties of the SW Conditions Corresponding to the GMCs and MS Intervals
ParameterMinMaxMedianMost ProbableDispersion
GMCMSGMCMSGMCMS
  • a

    In logarithmic scale.

δY, deg−1921−2.8−1.9−3.0−3.03.95.5
δZ, deg−17170.67−0.19−1−14.13.9
Psw, nPa3.7150202220230.25a0.25a
Bz, nT−6095−10−14−13−221112

[20] The doted histograms in Figures 5c and 5d correspond to common statistical distributions accumulated, respectively, for the solar wind pressure (1-min time resolution) and IMF Bz (16-s time resolution) measured by ACE in 1998 to 2001. From Figures 5c and 5d one can see that statistical distributions of the Psw and IMF Bz, corresponding to the MS intervals, are significantly different from the common distributions. The GMCs occur mostly under strongly disturbed solar wind conditions, which are associated with far tail of the common statistical distributions: Psw ∼ 20 nPa and Bz ∼ −14 nT. Sometime the GMCs can also be observed when the SW pressure is moderate (∼5 nPa) but the IMF Bz is negative and strong, or vice versa, the IMF is northward but the Psw is very high.

[21] Figure 6 represents scatter plots of the IMF components (in aGSM), which accompany the MS intervals. In Figures 6a and 6c the common statistical distributions are shown by dotted histograms, respectively, for the IMF Bx and By (16-s time resolution) measured by the ACE in 1998 to 2001. We have to emphasize that the SW conditions accompanying the MS intervals (solid histograms) are only a subset of the common distribution, which is “filtered” by a requirement to provide the magnetosheath location at geosynchronous orbit. The slopes of fitting of the scatter plots with a linear function are small, and thus the IMF components are independent. The scatter plot of By versus Bx (Figure 6a) demonstrates no evidence of the IMF orientation along the Parker spiral. Moreover the slope of the dependence of the By from Bx is positive and much different than visual orientation of the scatter plot. The above-mentioned features testify to practically random orientation of the IMF projection to aGSM equatorial plane.

image

Figure 6. Scatter plots of the IMF components (in aGSM), which accompany the MS intervals: (a) By versus Bx, (b) Bz versus Bx, and (c) Bz versus By. The dashed-dotted line in panel (c) indicates average direction of the Parker spiral. The best fit to the scatter plots with a liner functions is indicated by the dashed lines. The dotted histograms indicate common statistical distributions for the IMF (Figure 6a) Bx and (Figure 6c) By (in GSM) measured by the ACE in 1998 to 2001 (16-s time resolution). The solid histograms represent statistical distributions of the IMF Bx and By, which accompany the MS intervals.

Download figure to PowerPoint

[22] As a rule, the By and Bz components are larger than the Bx that indicates to relatively small influence of the IMF Bx component to the earthward motion of the dayside MP. The scatter plot of the Bz versus By (Figure 6c) has very wide spreading and practically circular shape. Hence the MS intervals are accompanied usually by the large Bz and/or By components. The IMF By scattering is symmetrical relative to the origin that supports the previous results obtained by Dmitriev and Suvorova [2000]. They reveal that the dayside MP dynamics does not depend on sign of the By, but the magnetopause size decreases in response to an increase in the IMF By absolute value. The Bz scattering is centered near Bz ∼ −13 nT as shown also on the statistical distribution presented in Figure 5d. The shift of the most probable Bz toward the negative values is explained apparently by the well-known fact that the IMF turning southward causes earthward magnetopause motion.

[23] Figure 7 displays the two-dimensional distribution of the occurrence number of the IMF orientation in the coordinate space of the clock angle versus the azimuth angle. The distribution is based on the SW data corresponding to the MS intervals. The clock angle is calculated from the aGSM equatorial plane, such that it is positive/negative for northward/southward IMF. The azimuth angle is calculated from the X-axis and it reflects the direction of the IMF projection in the aGSM equatorial plane. The azimuth angle of −45° corresponds to the IMF, which is aligned with the Parker spiral (indicated by the thick dashed line). The occurrence probability for the observing MS interval decreases quickly with the clock angle rotating from the negative to positive values, i.e., when the IMF Bz turns northward. This is in good agreement with the exponential decrease of the occurrence probability for observing very high SW pressure (see Figure 5c), which is required to push the magnetopause at the geosynchronous orbit under small negative and positive Bz. From Figure 7 we can see that the maximum of the occurrence number is located mostly in the range of an angle >60°, both for the clock angle and for the azimuth angle. A region of 30° in the central part of the two-dimensional distribution, corresponding to small Bz and By, relative to Bx, is practically empty. Hence the magnetosheath intervals are mostly accompanied by a randomly rotated large tangential component of the IMF.

image

Figure 7. Two-dimensional occurrence number distribution of the IMF orientation in coordinate space of the clock angle versus the azimuth angle that accompanies the MS intervals. The azimuth angle of −45°, corresponding to the Parker spiral, is indicated by the thick dashed line.

Download figure to PowerPoint

[24] In Figures 6 and 7 one can find no evidence for the Parker spiral angle appears in the IMF. The Parker spiral was demonstrated in many comprehensive statistical studies of the solar wind plasma and IMF conditions [e.g., Luhmann et al., 1993]. Detail consideration of the occurrence of various spiral angles in the IMP 8 data set (Figure 8b of Luhmann et al. [1993]) reveals significant contribution (∼20% of the maximum occurrence) from the spiral angles of −45° and 135°, which are perpendicular to the Parker spiral. That angles can be attributed both to the “noise” in the IMF and to specific SW structures such as magnetic clouds. In the latter case the vector of magnetic field rotates gradually and randomly in respect to the Sun-Earth line. The GMC events are caused by strongly disturbed solar wind conditions, especially when the IMF Bz is negative and large. These conditions correspond to the far tail of the common statistical distributions, and thus their properties might be different from the nominal SW. Indeed, strongly disturbed solar wind conditions are associated mostly with stream interaction regions, magnetic clouds, and the draped fields around them [e.g., Luhmann et al., 1993]. Hence the IMF orientation during the MS intervals is not necessarily aligned with the Parker spiral.

[25] The Psw and Bz are independent parameters; therefore their influence on the magnetopause is owed to different physical processes. The GMCs are caused by a superposition of two independent and intensively operating effects: magnetosphere compression by the SW pressure enhancement and/or magnetopause erosion under negative Bz. The dawn-dusk asymmetry is probably affected by the Psw and Bz in a different manner. We analyze the solar wind conditions for GMCs and magnetosheath intervals for their dependence on local time to study the role of the Psw and Bz in the dawn-dusk asymmetry.

[26] Scatter plots of the SW total pressure versus local time LT are presented in Figure 8a for Bz > 5 nT, Figure 8b for Bz > −15 nT, and Figure 8c for Bz > −60 nT. The GMCs and MS intervals observed by the geosynchronous satellites are indicated by crosses and triangles, respectively. The solar wind pressure varies very widely for the GMCs. Owing to magnetopause flaring, higher pressure is required to produce crossings on the flanks. For large positive IMF Bz (Figure 8a) the GMCs are caused by only the SW pressure enhancements, when Psw > 15 nPa. The scatter plot of Psw versus LT is practically symmetric. The GMCs and magnetosheath intervals associated with the minimal SW pressure Psw ∼ 15 nPa are observed close to noon, while a very high pressure Psw ∼ 70 nPa is observed for GMCs at 0700 LT or at 1700 LT. Note that previous studies of the GMCs could not reveal the symmetry in the distribution of the SW pressure versus the LT, owing to limited positive IMF Bz statistics.

image

Figure 8. Scatter plots of the solar wind total pressure versus local time for the GMCs collected from the GOES and LANL magnetopause crossings (crosses) and MS intervals (triangles) for (a) Bz > 5 nT, (b) Bz > −15 nT, and (c) all the data (Bz > −60 nT). The dashed curve is an approximation of the envelope boundary obtained for Bz < −6 nT by Kuznetsov and Suvorova [1997, 1998a].

Download figure to PowerPoint

[27] For negative IMF Bz (Figures 8b and 8c) the situation changes dramatically. For comparison with previous studies we reproduce in Figures 8b and 8c a boundary (dashed curve) enveloping the minimal dynamic pressures required for the GMCs obtained by Kuznetsov and Suvorova [1997, 1998a], for Bz < −6 nT. The boundary approximation was performed on the basis of a smaller data set containing only 84 GMCs detected in 1967 to 1993. As we can see, the previous approximation is in good agreement with our results for Bz > −15 nT. The scatter plots of Psw versus LT demonstrate very strong dawn-dusk asymmetry for the large southward IMF. The minimal SW pressures Psw ∼ 5 nPa are observed mostly in the prenoon sector at ∼1000 LT. In Figure 4b one can see that at 0800 LT, the SW pressure Psw ∼ 7 nPa is enough to produce the GMC, while at 1600 LT a SW pressure of Psw ∼ 14 nPa (2 times more) is required. The asymmetry becomes larger for a larger negative Bz. From Figure 8c we can estimate that the minimal SW pressure required for the GMC is about 10 nPa at 0700 LT and Psw ∼ 30 nPa at 1700 LT. Hence the difference of the SW pressures required for GMCs on the dawn and dusk flanks is as much as 3 times or more. Such a difference is independent evidence of dawn-dusk asymmetry of the GMCs under a strong negative IMF Bz.

5. Applications for a MP Model

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Data
  5. 3. LT Distribution of the GMCs
  6. 4. Dependence on the Solar Wind Conditions
  7. 5. Applications for a MP Model
  8. 6. Discussion
  9. 7. Conclusions
  10. Acknowledgments
  11. References

[28] Previous MP models [Kuznetsov and Suvorova, 1997; Dmitriev and Suvorova, 2000] include the MP asymmetry in a different and sometimes complicated manner. We introduce here the derivation of the dawn-dusk asymmetry using the Chao et al. [2002] MP model (CH01 model) with a simple conversion procedure of the model coordinate system by rotation around the GSM Z-axis at some arbitrary angle. In this case the dayside magnetopause is easily described as a revolution surface with a new X′-axis, which is rotated from the Earth-Sun axis X in the GSM equatorial plane. Apparently, this simple method is only the first mathematical approach to the actual shape of the dayside magnetopause under strongly disturbed solar wind conditions.

[29] The best-fit technique is used to determine the rotation angle value, which provides the best accuracy of the model for the GMC prediction. The model accuracy is determined as a standard deviation of the magnetopause distance predicted by the model from the geosynchronous orbit distance (6.6. RE). The initial version of the CH01 model has a standard deviation of SD = 0.55 RE for the prediction of the GMCs. The application of the best-fit technique permits us to obtain an angle of the magnetopause axis rotation equal to θ = 15° toward the dawn. The modified version, with rotated axis of revolution gives a much more accurate GMC prediction, with SD = 0.45 RE. Therefore the simple conversion of the coordinate system, with allowance for the effect of the 15° dawn-dusk asymmetry, permits a significant (about 20%) improvement in the MP model capability in predicting geosynchronous magnetopause crossings.

[30] This improvement is demonstrated in Figure 9 for the case event of 6–7 April 2000. The prediction of the MP location from the initial CH01 model is shown on the top panel. The bottom panel shows the prediction of the modified version of the CH01 model. The horizontal dashed lines indicate the distance to the geosynchronous orbit (6.6 RE). The ordinate axis shows the deviation from the geosynchronous orbit in the Earth's radii. The abscissa is the universal time (UT). The model predictions are compared with the GOES-8 magnetosheath intervals (hatched bars) when the magnetopause is located inside the geosynchronous orbit (negative deviation). One can clearly see that the accuracy of the modified CH01 model, for the prediction of the magnetopause location, becomes much higher. Indeed, the initial CH01 model incorrectly predicts the 1840 UT to 1930 UT magnetosheath interval. The modified version of the model successfully rejects this false interval.

image

Figure 9. Comparison of predictions of the magnetopause location near geosynchronous orbit (dashed line) from the CH01 model, without (top) and with (bottom), introducing the dawn-dusk MP asymmetry. The hatched bars are GMC events observed by GOES-8.

Download figure to PowerPoint

[31] An estimation of the model accuracy by the probability of correct prediction (PCP, PoP), probability of detection (PoD) and false alarm rate (FAR) [Shue et al., 2000; Yang et al., 2002] shows that the modified CH01 model has the same PoP and PoD but a much lower FAR (about 50% less) for the 6–7 April 2000 event, when the duration of the GOES-8 magnetosheath interval is about 40 min. Hence the CH01 model, including the dawn-dusk asymmetry, is more capable of predicting the GMC events. The recent study devoted to detail comparison of different magnetopause models under extreme SW condition with in situ observations [Shue et al., 2001] reveals that the Kuznetsov and Suvorova [1998b] magnetopause model is the best one (because the highest PCP) to predict the magnetosheath periods under extremely strong southward IMF. As we indicated above, the Kuznetsov and Suvorova [1998b] model represents the dawn-dusk asymmetry under strong negative Bz. Hence an incorporation of the dawn-dusk MP asymmetry into the models increases their capability to predict the geosynchronous magnetopause crossings especially under southward IMF.

6. Discussion

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Data
  5. 3. LT Distribution of the GMCs
  6. 4. Dependence on the Solar Wind Conditions
  7. 5. Applications for a MP Model
  8. 6. Discussion
  9. 7. Conclusions
  10. Acknowledgments
  11. References

[32] In the present study we have collected 515 magnetopause crossings by geosynchronous satellites using literature sources (189 GMCs), magnetic measurements on GOES satellites (129 GMCs), and plasma measurements on LANL satellites (197 GMCs). Unfortunately, the effects of selection cannot be completely eliminated from the GMC data set because precise calculation of the observation time for each LT is practically impossible. However, the great amount of the data accumulated from 1967 to 2001 should almost completely eliminate this effect. Moreover, an independent evidence of the dawn-dusk asymmetry is that the scatter plot of Psw versus LT does not depend on the effect of data accumulation. The other important property of the selected GMC data set is that it is highly inhomogeneous in terms of methods of GMC identification and determination of the corresponding solar wind conditions. The GMCs were identified by different authors and different methods are used to analyze the magnetic field and plasma data obtained from different geosynchronous satellites. Despite the difference in methods, all the considered GMC sets are in good agreement with one another, as we have shown in section 3. Hence we can conclude that experimental methods of GMC identification and determination of the SW conditions do not influence the basic properties of the collected GMC data set. The obtained dawn-dusk asymmetry of the GMCs cannot have originated from the artificial effects of selection. It is also free from any systematical errors. Thus we can suggest that the local time asymmetry of the GMCs is associated with the dawn-dusk asymmetry of the magnetopause.

[33] We show that the magnetopause is not always asymmetrical but only for a strong negative Bz. This is one of the reasons for the contradictory results obtained in previous studies. The other apparent reason is the limitation of the GMC statistics. It was impossible to obtain statistically significant evidence of the asymmetry in the past. The third reason is the different experimental methods used for GMC data treatment in previous studies. This difference obviously leads to inconsistencies in results by different investigators. A statistical analysis performed for magnetosheath intervals provides significantly different results than one performed for magnetosheath entrances. Indeed, in the case of magnetosheath intervals, the GMC has a contribution proportional to its duration such that the bins in the vicinity of noon accumulate numerous redundant statistics, owing to crossings actually occurred far from noon. Apparently the nearness to noon, for the median and mean values in the McComas et al. [1994] histogram can be explained by redundant statistics that have been accumulated in the noon region. Therefore the method of the magnetosheath interval occurrence number is not suitable for the study of the LT asymmetry of the GMC occurrence probability. On the other hand, a magnetopause crossing itself directly indicates the MP location for given solar wind conditions. Thus the GMC occurrence probability at different local times directly reflects the properties of the dayside magnetopause shape. In other words, the magnetopause asymmetry can be verified statistically from an analysis of the local time distribution of the GMC occurrence probability. This suggestion is incorrect in the case of long-lasting GMCs. However, statistical analysis of the GMC duration (Figure 2) shows that most GMCs are shorter than dT ∼ 20 min, and thus we can neglect the GMC duration effect.

[34] The magnetopause skewing toward dawn under a southward IMF was first reported by Rufenach et al. [1989]. We find that the maximal probability of magnetosheath encounters in geosynchronous orbit is observed between 1000 LT and 1100 LT. On the other hand, GMCs are mostly observed when the IMF Bz is negative. In this sense our result is in good agreement with Rufenach et al.'s [1989]. The maximum of the GMC occurrence numbers at 1100 LT means that the nearest to the Earth region of the magnetopause has shifted by 1 hour from noon toward the morning. The shift can be interpreted in four ways. The first one is that the magnetopause rotates about Z-GSM axis at an angle of θ = 15° toward the dawn. The second interpretation is that the magnetopause fluctuations are much larger in the prenoon sector, especially under a southward IMF. The third interpretation is that the magnetopause erosion is much more intensive in the prenoon sector than in the postnoon sector. The fourth interpretation is that the magnetopause axis shifts from the X-GSM axis toward dusk on a few RE.

[35] The rotation of the magnetopause nose point toward dawn was predicted in early theoretical studies of solar wind flow around the magnetopause [Walters, 1964]. It was shown that for oblique IMF oriented along the Parker spiral, the jump conditions across the terrestrial bow shock are weaker on the dusk side, where the shock is mostly quasi-perpendicular, than on the dawn side, where the quasi-parallel shock prevails. Owing to this asymmetry in bow shock formation, the stagnation point on the magnetopause should be shifted toward dawn at an angle from θ = 4° to ∼20°, depending on the solar wind plasma and the IMF properties. For nominal solar wind conditions this effect is small and θ ∼ 4°. When the IMF is strong and southward, Walters's [1964] theory predicts that the magnetopause should have a north-south rather than a dawn-dusk asymmetry, especially for a relatively low solar wind pressure. This effect is “perpendicular” to our results and hence does not help their interpretation.

[36] The dawn-dusk asymmetry for the occurrence probability of the MP crossings and the amplitude of the magnetopause oscillation was found by Russell et al. [1997] for an IMF aligned with the Parker spiral, which is mostly observed under typical solar wind conditions. It is suggested that a foreshock effect is responsible for the relatively large magnetopause oscillations in the prenoon sector, especially under a southward IMF. Owing to the large amplitude oscillations, the magnetopause crossings are observed more often in the prenoon sector, where the foreshock is usually strong. On the other hand, we have shown (Figures 6 and 7) that there is no preferred IMF direction associated with the Parker spiral under the strongly disturbed solar wind conditions, which accompany the GMCs. Hence the region with the strongest foreshock should be observed in both the prenoon and the postnoon sectors with practically equal probabilities. Note that in the case of the upstream solar wind structures such as foreshocks, corotating discontinuities, etc., the dawn-dusk magnetopause asymmetry is controlled by the IMF orientation in respect to the Parker spiral [e.g., Song, 1994]. Therefore it is unlikely that the foreshock effect as well as other upstream solar wind phenomena contribute to the dawn-dusk asymmetry of the GMCs.

[37] Discussing the dawn-dusk magnetopause asymmetry, McComas et al. [1993] has suggested that the prenoon magnetopause should approach the Earth owing to more intensive erosion on the dawnside than on the duskside. The reasons for this erosion asymmetry were unclear. This suggestion is supported by our conclusion that the dawn-dusk asymmetry can be observed under a southward IMF, which causes magnetopause erosion. Additional support of this idea can be found in a theoretical study of reconnection at the MP [Dreher and Schindler, 1997]. It is found that in the early phase of the reconnection process, when Bz is negative, the Hall term results in a systematic dawnward transport of the magnetic flux. Dreher and Schindler conclude that this effect may account for dawn-dusk asymmetries in the occurrence of the flux transfer events as well as for other magnetopause phenomena. Unfortunately, the significance of this effect cannot be estimated numerically from the theory.

[38] Modeling the dawn-dusk magnetopause asymmetry under strongly disturbed SW conditions, Kuznetsov and Suvorova [1997] proposed a shift of the magnetopause axis from the X-GSM axis toward dusk of up to 2 RE. Physically, this idea means that the magnetopause on the dusk side has an additional internal source of energy, which produces stronger resistance of the magnetopause to the solar wind pressure. From Figure 8 we can estimate the difference between the internal magnetosphere energy at dawn and at dusk required for the observed dawn-dusk asymmetry. Taking the solar wind pressure Psw = 30 nPa at dusk and Psw = 10 nPa at dawn, we obtain the ratio Pswdusk/Pswdawn = 3. An internal energy source that provides such strong asymmetry could be the well-known ring current, developed under the negative IMF Bz, which has an asymmetrical local time distribution with the maximum in the dusk sector. From the pressure balance, we can estimate the ring current asymmetry, providing the magnetic effect required for the magnetopause asymmetry. A solar wind pressure ratio of 3 is equivalent to the ratio of the magnetospheric field Hdusk/Hdawn = √3 = 1.7. The background geomagnetic field at geosynchronous orbit is about H = 100 nT so the additional contribution of the ring current magnetic effect in the geosynchronous magnetic field in the dusk sector should be more than ∼70 nT. Such a strong magnetic effect of the ring current takes place only during severe geomagnetic storms, which are caused by strong negative IMF Bz. This supports our results on the magnetopause dawn-dusk asymmetry.

[39] Another possible source of the additional magnetosphere energy first proposed by Suvorova et al. (submitted manuscript, 2004) to explain the MP dawn-dusk asymmetry is the thermal pressure of the magnetosphere plasma. The estimation of the perpendicular thermal pressure at dawn (0600 LT) and at dusk (1800 LT) [Michelis et al., 1999] shows that during disturbed geomagnetic activity, the dusk to dawn pressure ratio in geosynchronous orbit is about 1.3. Apparently, such a ratio can only partially contribute to the observed asymmetry. Hence both the magnetospheric thermal pressure and the magnetic effect should be accounted together as a manifestation of the storm associated asymmetrical ring current.

[40] Our consideration of the different physical processes responsible for the magnetopause dawn-dusk asymmetry at geosynchronous orbit permits the revealing of the two most probable driving processes: intensive magnetopause erosion in the prenoon sector and strong asymmetry of the terrestrial ring current during geomagnetic storms. The contribution of these mechanisms is still beyond precise numerical calculation. In such a situation the modeling of the magnetopause dawn-dusk asymmetry should be a very useful way to describe the response of the magnetosphere to strongly disturbed solar wind conditions.

7. Conclusions

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Data
  5. 3. LT Distribution of the GMCs
  6. 4. Dependence on the Solar Wind Conditions
  7. 5. Applications for a MP Model
  8. 6. Discussion
  9. 7. Conclusions
  10. Acknowledgments
  11. References

[41] 1. A substantial dawn-dusk GMC asymmetry is obtained using two independent methods for 515 GMCs and magnetosheath intervals collected from the literature sources and selected from magnetic measurements on the GOES satellite and plasma measurements on the LANL satellites.

[42] 2. The LT distribution of the GMC occurrence probability reveals that the GMCs are mostly observed in the local time range from 1000 LT to 1100 LT.

[43] 3. For the MS intervals an analysis of the scatter plot of the solar wind pressure versus the local time shows convincingly that for a large positive IMF Bz, the magnetopause is symmetrical. A prominent magnetopause dawn-dusk asymmetry is developed under strong negative Bz. The asymmetry is characterized by shifting the GMCs, with the minimal required total solar wind pressures, toward dawn. The solar wind pressures providing the GMCs in the dawn sector are significantly lower (about 3 times) than in the dusk sector.

[44] 4. The magnetopause dawn-dusk asymmetry application for a magnetopause model permits a substantial increase in the model's capability for the prediction of the geosynchronous magnetopause crossings. The standard deviation of the Chao et al. [2002] model decreases by 20%, from 0.55 RE to 0.45 RE.

[45] 5. The most probable physical phenomena responsible for the magnetopause dawn-dusk asymmetry are the more intensive magnetopause erosion in the prenoon sector and strong asymmetry of the terrestrial ring current during severe geomagnetic storms.

Acknowledgments

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Data
  5. 3. LT Distribution of the GMCs
  6. 4. Dependence on the Solar Wind Conditions
  7. 5. Applications for a MP Model
  8. 6. Discussion
  9. 7. Conclusions
  10. Acknowledgments
  11. References

[46] The authors thank NASA/GSFC ISTP for providing data from the Geotail, Wind, ACE, GOES, and LANL satellites. We thank T. Nagai from Tokyo Institute of Technology Earth and Planetary Sciences for providing the Geotail magnetic data, L. A. Frank from the University of Iowa for providing the Geotail plasma data, and R. P. Lepping from NASA/GSFC for providing the Wind magnetic data. We also thank collaborative efforts of GSFC, UNH, and MIT in providing the Wind plasma data, C.W. Smith from University of New Hampshire for providing the ACE magnetic data, Dave J. McComas from Los Alamos National Laboratory for providing the ACE plasma data, NASA and NOAA for providing the GOES magnetic data, Los Alamos National Laboratory for providing the LANL plasma data, and Kyoto World Data Center for Geomagnetism for providing the ASY/SYM indices. This work was supported by grant NSC-92-2111-M-008-003.

[47] Lou-Chuang Lee thanks the two reviewers for their assistance in evaluating this paper.

References

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Data
  5. 3. LT Distribution of the GMCs
  6. 4. Dependence on the Solar Wind Conditions
  7. 5. Applications for a MP Model
  8. 6. Discussion
  9. 7. Conclusions
  10. Acknowledgments
  11. References