SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Observations
  5. 3. Summary and Discussion
  6. Acknowledgments
  7. References

[1] “Magnetic clouds” (MCs) are a subset of interplanetary coronal mass ejections (ICMEs) characterized by enhanced magnetic fields with an organized rotation in direction, and low plasma β. Though intensely studied, MCs only constitute a fraction of all ICMEs detected in the solar wind. A comprehensive survey of ICMEs in the near-Earth solar wind during the ascending, maximum and early declining phases of solar cycle 23 in 1996–2003 shows that the MC fraction varied with the phase of the solar cycle, from ∼100% (though with low statistics) at solar minimum to ∼15% at solar maximum. A similar trend is evident in near-Earth observations during solar cycles 20–21, while Helios 1/2 observations at 0.3–1.0 AU show a weaker trend and larger MC fraction.

1. Introduction

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Observations
  5. 3. Summary and Discussion
  6. Acknowledgments
  7. References

[2] Interplanetary coronal mass ejections (ICMEs), the manifestations in the solar wind of coronal mass ejections (CMEs) at the Sun, are characterized by various signatures [e.g., Gosling, 1990; Zurbuchen and Richardson, 2004] that include abnormally low plasma proton temperatures, bidirectional suprathermal electron strahls (BDEs) and energetic particle flows, cosmic ray depressions and plasma compositional anomalies. Fast ICMEs also generate shocks in the upstream solar wind. Klein and Burlaga [1982] defined a specific subset of ICMEs with enhanced magnetic field strengths >10 nT, a smooth rotation of the magnetic field direction through a large angle, durations of ∼1 day, and low plasma β (ratio of plasma/magnetic field pressures). Such “magnetic clouds” (MCs) have been the focus of intense study that includes modeling as force-free [e.g., Lepping et al., 1990] or non-force-free flux ropes [e.g., Osherovich and Burlaga, 1997; Cid et al., 2002]. Flux rope configurations may arise naturally during CME eruptions [e.g., Gosling et al., 1995], and are suggested by helical structures visible in coronagraph images of CMEs [e.g., Chen et al., 1997; Dere et al., 1999]. Magnetic clouds with periods of strong southward magnetic field are also responsible for many intense geomagnetic storms [e.g., Cane et al., 2000; Webb et al., 2000; Cane and Richardson, 2003, and references therein]. The large-scale structure of MCs has been examined using observations from multiple spacecraft [e.g., Burlaga et al., 1981; Cane et al., 1997; Mulligan et al., 1999a]. Figure 1a shows solar wind magnetic field and plasma observations from the Advanced Composition Explorer (ACE) during passage of an ICME on April 16–17, 1999, bounded by the vertical dashed lines, that exhibits the characteristic features of an MC as discussed above.

image

Figure 1. ACE solar wind magnetic field and plasma parameters during ICMEs with (a) a clear MC signature, (b) a weak, fluctuating magnetic field with no obvious organization, and (c) a weak field with two intervals of organized rotations in direction. All follow a shock (solid vertical line) and are associated with regions of depressed proton temperatures (black shading).

Download figure to PowerPoint

[3] Nonetheless, many ICMEs lack the characteristics of MCs. Figure 1b shows an example on March 10–12, 1999. The magnetic field is <10 nT, includes several directional discontinuities and has no large-scale organization. An intermediate event (October 5–7, 2000; Figure 1c) includes two periods of reasonably organized field rotations (delineated by vertical dashed lines) though the field intensity is <10 nT. Otherwise, the ICMEs in Figure 1 share characteristics such as periods of abnormally low proton temperatures relative to the “expected” temperature [e.g., Richardson et al., 1997] indicated by black shading, declining speed profiles consistent with expansion, and upstream shocks (solid vertical lines). Zurbuchen and Richardson [2004] illustrate additional ICME signatures for these events.

[4] Various studies have estimated the fraction of ICMEs that are MCs. Gosling [1990] concluded that ∼30% of ICMEs in 1978–1982 defined by BDEs were MCs, a widely quoted result. Mulligan et al. [1999b] obtained 38% for ICMEs in mid-1978–1979 identified using several signatures but with a more relaxed MC definition, while only 14% of the ICMEs in 1978–1982 discussed by Richardson et al. [1997] were MCs. Cane et al. [1996] found that ∼49% of the ICMEs associated with >3% cosmic ray decreases at Earth in 1964–1994 were MCs. Bothmer and Schwenn [1996] concluded that ∼41% of ICMEs encountered by the Helios 1 or 2 spacecraft at 0.3–1.0 AU in 1979–1981 were MCs (increasing to ∼60% for their events that followed strong shocks [Cane et al., 1997]). An MC fraction of ∼60% was estimated by Cane et al. [1997] for Helios 1/2 ICMEs in 1975–1979 associated with cosmic ray depressions, while Marubashi [2000] has claimed that up to ∼80% of near-Earth ICMEs during 1978–1982 were magnetic flux rope encounters. Henke et al. [2001] found a 32% MC fraction for Ulysses ICMEs in 1991–1996. In this paper, we point out that the fraction of MCs appears to depend on the phase of the solar cycle, both in the current and earlier cycles.

2. Observations

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Observations
  5. 3. Summary and Discussion
  6. Acknowledgments
  7. References

[5] When assessing the fraction of ICMEs that are MCs, a comprehensive list of ICMEs is required to provide a reliable normalization for the number of MCs. The present study uses first a list of near-Earth ICMEs compiled for the years 1996–2002, encompassing the rising and maximum phases of solar cycle 23 [Cane and Richardson, 2003]. This list, based on in-situ observations (principally of solar wind plasma and magnetic field, which are essentially complete for this solar cycle), aims to provide a comprehensive survey of ICMEs in the near-Earth solar wind. We have updated the list to include events in 2003 and a few minor corrections. A recent study [Richardson and Cane, 2004] indicates that solar wind compositional anomalies characteristic of ICMEs show an excellent association with the Cane and Richardson [2003] ICMEs.

[6] Figure 2a shows the monthly sunspot number during 1996–2003, from the Royal Observatory of Belgium. Reflecting the increase in the CME rate at the Sun observed by LASCO [St. Cyr et al., 2000], the ICME rate during this period (Figure 2b) increases by around an order of magnitude, from 4/year in 1996, to ∼50/year around solar maximum (the lower rate in 1999 was noted by Cane et al. [2000]). To identify ICMEs that are MCs, we have referred to the MC list compiled by the WIND MFI team (http://lepmfi.gsfc.nasa.gov/mfi/mag_cloud_pub1.html). We generally corroborate these identifications, though with occasional exceptions. Figure 2b indicates the number/year of our ICMEs that are MCs, while Figure 2c shows the ratio of MCs to all ICMEs. This ratio clearly varies with the solar cycle, from 100% in 1996 (with low statistics) to ∼15% around solar maximum. There is evidence of a recovery in 2002 as the ICME rate declines, but this does not appear to have continued into 2003. “Error bars” in Figure 2c (and Figures 3 and 4 below) indicate how the MC fraction would change if the number of ICMEs or MCs were changed by one event.

image

Figure 2. (a) Monthly sunspot number for 1996–2003; (b) Total number of ICMEs/year, updated from Richardson and Cane [2003] and the number of these events that are magnetic clouds; (c) Percentage of ICMEs that are magnetic clouds.

Download figure to PowerPoint

image

Figure 3. Variation in the MC fraction vs. years from sunspot minimum for Cane et al. [1996] class 1 or 3 events (•) and ICMEs in 1972–1982 (cycles 20–21; ○).

Download figure to PowerPoint

image

Figure 4. Variation in the MC fraction for Helios 1/2 ICMEs vs. time from sunspot minimum for heliocentric distances 0.3–1.0 AU and 0.3–0.8 AU.

Download figure to PowerPoint

[7] We next examine whether previous solar cycles show a similar variation in the MC fraction. Filled circles in Figure 3 show the MC fraction as a function of time relative to solar minimum for events during 1964–1994 in the cosmic ray study of Cane et al. [1996] where an ICME was present (“Class 1/3” events). Only the 76 events with adequate magnetic field coverage are included. Though again there are few ICMEs near solar minimum (years from minimum with ≤2 events are not shown in the figure), there is evidence of a decline in the MC fraction from ∼40% during 1–3 years after minimum to ∼25% during 3–7 years after minimum (i.e., around solar maximum). Also shown in Figure 3 by open circles is the MC fraction for 204 ICMEs with magnetic field data available that we have identified in the OMNI solar wind data for 1972–1982, encompassing the decline of solar cycle 20 to the decline of cycle 21 (the most extended period of nearly complete 1 AU solar wind observations prior to the current cycle). Again, the MC fraction decreases as solar activity levels increase, then recovers as the cycle declines, with the notable exception of the low MC fraction 1–2 years before sunspot minimum, which may be associated with a temporary increase in solar activity during 1974.

[8] We have also estimated the MC fraction for ∼150 probable ICMEs observed at 0.3–1.0 AU by Helios 1 or 2 in 1975–1981 or 1976–1980, respectively. These ICMEs are identified predominantly from plasma and magnetic field signatures, cosmic ray modulations [Cane et al., 1997], and associations with interplanetary shocks. The Helios MC fraction (Figure 4, filled circles) again tends to decline with time from solar minimum, though this variation is much weaker than that near the Earth. In addition, overall, the MC fraction is higher than that inferred from near-Earth studies [cf. Cane et al., 1997; Bothmer and Schwenn, 1996]. This result is not understood. One possibility is that the MC fraction may be higher closer to the Sun. However, grouping the Helios ICMEs into various radial distance ranges provides no clear evidence that this is the case. For example, open circles in Figure 4 show the MC fraction at 0.3–0.8 AU—the Helios spacecraft spent approximately equal periods inside and outside of 0.8 AU.

3. Summary and Discussion

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Observations
  5. 3. Summary and Discussion
  6. Acknowledgments
  7. References

[9] Observations near Earth during cycle 23 and cycles 20–21 indicate a decrease in the fraction of ICMEs that are magnetic clouds as solar activity levels increase, from ∼70% around solar minimum to ∼20% around solar maximum. This trend is also present, though much weaker, in Helios 1 and 2 observations at 0.3–1.0 AU from the Sun. However, the Helios observations also suggest higher MC fractions (∼60%) than those typically found near the Earth for reasons that are still unclear but are apparently unrelated to the spacecraft distance from the Sun.

[10] Several factors could contribute to the solar cycle variation in the MC fraction. One possibility is that MCs near the Sun evolve into more complicated structures by interactions with other ICMEs en route to 1 AU [e.g., Burlaga et al., 2002]. (Thus, the ICME in Figure 1c may include two components with different senses of field rotation.) Such interactions are more likely at high activity levels, consistent with the decrease in the MC fraction around solar maximum. However, to reduce the MC fraction from say 70% to ∼20% as a result of single ICME-ICME interactions would require ∼80% of ICMEs to make an interaction. This high prevalence of interactions seems unlikely given typical separations of ∼1 week between ICMEs passing the Earth at solar maximum (see Figure 2) which is considerably longer than typical ICME transit times to 1 AU. Furthermore, the absence of a clear decrease in the Helios MC fraction with heliocentric distance suggests that if such interactions do influence the MC fraction, they must predominantly occur inside the orbital range of Helios (<0.3 AU from the Sun).

[11] The increasing latitudinal spread of CME central axes as activity levels increase [Hundhausen, 1993; St. Cyr et al., 2000] may also increase the probability that an MC will only make a glancing encounter with the Earth or low-heliolatitude spacecraft. In such cases, clear MC signatures may not be observed, reducing the apparent MC fraction at low latitudes as activity increases. Indeed, Henke et al. [2001] reported a higher (∼50%) fraction for their high-latitude Ulysses ICMEs than for their low-latitude events. However, because of the spacecraft orbit, the high- or low-latitude ICMEs were observed during quiet or active conditions, respectively. Thus, the MC fraction may simply have reflected the solar cycle variation discussed in this paper, rather than a true latitudinal dependence.

[12] Our favored cause of the solar cycle variation is an increase in the typical magnetic complexity of CMEs as solar activity intensifies. Simple flux-rope like configurations may be predominant at low activity levels. As activity increases, reconnection of multiple loop systems, and the complex overlying coronal field, possibly involving structures far from the source region, may result in CMEs and ICMEs with more complicated magnetic field structures.

[13] A solar cycle variation in the MC fraction may have important implications for space weather forecasting. For example, the probability of a given ICME propagating towards the Earth producing a geomagnetic storm may be higher at solar minimum since there is a greater likelihood of it having an MC-like magnetic field which in turn may have a prolonged, strong southward component [cf. Cane and Richardson, 2003, Figure 6]. Evidence of such an effect is found by Zhao and Webb [2003]. They note that the solar minimum period in 1996–1997 when six apparently Earthward-directed CMEs were all followed by shocks, MCs and moderate geomagnetic storms [Webb et al., 2000] is exceptional, and that relatively fewer geomagnetic storms followed similar CMEs closer to solar maximum. They attribute this to the increased inclination of the streamer belt at solar maximum which increases the inclination of the axis of MCs ejected from the streamer belt, reducing their geoeffectiveness.

[14] In addition to the solar cycle dependence suggested here, differences in the MC fraction reported in individual studies may be attributable to differences in the set of ICMEs used to normalize the number of MCs and the criteria used to compile the MC list (e.g., whether MCs with significant data gaps are excluded, or the Klein and Burlaga [1982] criteria relaxed). Furthermore, around half of 1 AU ICMEs show evidence of a field rotation that might be a relic of a cloud-like structure [Cane and Richardson, 2003], even though the fraction of “classic” MCs is much smaller. On the other hand, there are complicated events (e.g., Figure 1b), which seem difficult to interpret as a conglomeration of simple MC-like structures or a glancing encounter with a region of organized magnetic field. A final complication is that multi-spacecraft observations suggest that MCs can be substructures of ICMEs [Cane et al., 1997], so that the same ICME may be identified as an MC at one location but not at another.

Acknowledgments

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Observations
  5. 3. Summary and Discussion
  6. Acknowledgments
  7. References

[15] We gratefully acknowledge use of magnetic field and solar wind plasma data from the ACE, WIND, IMP 8 and Helios 1/2 spacecraft, provided via the ACE Science Center and NSSDC. Additional Helios data were provided by R. Schwenn. IGR is supported by NASA grant NCC 5-180 and HVC by a NASA contract with USRA.

References

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Observations
  5. 3. Summary and Discussion
  6. Acknowledgments
  7. References