- Email: [email protected]
Solar proton events during the solar cycle 23 and their association with CME parameters
ARTICLE IN PRESS Acta Astronautica 67 (2010) 353–361
Contents lists available at ScienceDirect
Acta Astronautica journal homepage: www.elsevier.com/locate/actaastro
Solar proton events during the solar cycle 23 and their association with CME parameters$ Junga Hwang a,, Kyung-Suk Cho a, Young-Jae Moon b, Rok-Soon Kim c, Young-Deuk Park a a b c
Korea Astronomy and Space Science Institute, 61-1, Hwaam dong, Yuseong gu Daejeon, 305-348, South Korea School of Space Research, Kyung Hee University, Yongin-si, Korea Chungnam National University, Daejeon, Korea
a r t i c l e in fo
abstract
Article history: Received 22 January 2010 Received in revised form 11 March 2010 Accepted 5 April 2010 Available online 28 April 2010
We have studied the solar proton events associated with the coronal mass ejections (CMEs) and flares during the solar cycle 23 (1997–2006) in order to determine what physical parameters of the solar eruptions might control the SPE intensity and time profile. For total 63 SPEs, we found that (1) SPE rise time, duration time and decrease times depend on a CME speed (cc= 0.34, 0.48 and 0.48) and (2) a SPE peak intensity depends on an earthward direction parameter of a CME as well as the CME speed and xray flare intensity (cc =0.40, 0.31 and 0.37). The SPEs were divided into two groups according to the correlation between the CME earthward direction parameter and the SPE intensity. First group consists of large six SPEs (>10,000 pfu at >10 MeV proton channel of GOES satellite) and shows a very good correlation (cc= 0.65) between the SPE peak intensity and the CME earthward direction parameter. Second group has a relatively weak SPE peak intensity and shows no correlation (cc= 0.01) between the SPE peak intensity and the CME earthward direction parameter we found that the first group SPEs are associated with a very fast halo CME (>1400 km/s) and most of those are located at disk except for only one case. Especially, large six SPEs have a good correlation with their associated CME earthward direction parameters, implying that these events are produced by ICME-driven shocks. We also found that those six SPEs are associated with the preceding CMEs originated from the same solar source region and a nearby pre-existing helmet streamer. Thus, we speculate that the preceding CME and helmet streamer might provide seed particles for CME-driven shocks and cause a clear separation between two groups. & 2010 Elsevier Ltd. All rights reserved.
Keywords: Solar energetic particle Corona mass ejection Solar flare
1. Introduction Solar flares and coronal mass ejections (CMEs) are the most powerful events in the solar system. In tens of minutes they can convert in excess of 1032 ergs of magnetic energy into accelerated particles, heated plasma, and ejected solar
$
This paper was presented during the 60th IAC in Daejeon.
Corresponding author. Tel.: + 82 42 865 2061; fax: + 82 42 865 2020.
E-mail addresses: [email protected], [email protected] (J. Hwang), [email protected] (K.-S. Cho), [email protected] (Y.-J. Moon), [email protected] (R.-S. Kim), [email protected] (Y.-D. Park). 0094-5765/$ - see front matter & 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.actaastro.2010.04.001
material. In this paper, we chose the word of ‘‘Solar Proton Events (SPEs)’’ instead of ‘‘Solar Energetic Particles (SEPs)’’ as we analyzed only proton data. Some large SPEs had a rise time (from a threshold to the maximum intensity) within tens of minutes. This rapid rise time, coupled with the high intensity, means that if astronauts had been on EVA (Extra Vehicular Activity) on the Moon, there would be very little warning before the significant enhancement of the SPE flux. Events of this nature need to be taken into account in planning astronaut operations and shelters on the Moon and in interplanetary space, which is because the space radiation exposure to a human body and spacecraft can make severe
ARTICLE IN PRESS 354
J. Hwang et al. / Acta Astronautica 67 (2010) 353–361
radiation hazards. In addition, the rapid rise places extreme requirements on models that accelerate SPEs by CME-driven shocks, because protons with energies 41 GeV must have been accelerated within minutes in the very low corona. There is a substantial evidence that gradual SPEs observed at 1 AU result from an acceleration at coronal and interplanetary shocks driven by fast CMEs [1,2]. CME speeds have been found to be correlated with the peak intensities of the associated SPEs, but there is a very broad scatter. This weak correlation indicates a complexity in the SPE shock acceleration process that has not been fully understood. Recently, several attempts have been made to explore the effects that variation in the coronal magnetic and particle environment might have in the resulting SPE production by CME-driven shocks. Enhanced decametric– hectometric (DH) emission was observed during interactions of a fast CME with a preceding slower CME [3,4]. The enhanced DH emission was interpreted as evidence for strengthening of a pre-existing shock of the slower CME and for the formation of a new shock in the fast CME. The basic interpretation was that the efficiency of particle acceleration is somehow enhanced when the primary CME runs into regions of enhanced density due to preceding CMEs or streamers [5]. The first large statistical study of E420 MeV SPEs was probably that of van Hollebeke et al. [6]. Using 185 20 MeV oEo80 MeV SPEs from the Interplanetary Monitoring Platform (IMP) spacecraft, they focused on the early phases of SEP events and plotted times DT, from event onsets to maxima, as a function of solar source longitude. The time DT reached a minimum value of about 2.5 h around W501 and increased for larger magnetic angular separations between the Earth and the solar source longitude. Cane et al. [7] surveyed a larger sample of 235 SEP events and plotted as a function of solar source longitude the delay times from associated Ha flare maxima to peak SEP intensities, some of which were peaks at 1 AU shock passages. In each of three proton energy ranges spanning 1–100 MeV, those plots showed a pattern of delay times increasing from W901 to E901 source longitudes. A similar result was found for 112 E4 10 MeV proton events by Balch [8]. The decay phases of gradual SEP events are characterized by a spatial and temporal invariance of the energy spectra, whose onset is ordered by the location of the observer relative to the interplanetary shock [9,10]. Statistical analyses of E4 4 MeV proton event decay phases show a tendency for event decay times to decrease with higher shock speeds and with steeper energy spectra [11], supporting the interpretation of adiabatic deceleration of SEPs quasitrapped behind large-scale expanding shocks. Since CMEs are the drivers of shocks that accelerate SEPs, we might expect that the characteristics of the SEP intensity–time profiles observed at 1 AU are determined by properties of the associated CMEs. For example, faster CMEs may drive shocks for longer periods of time, resulting in SEP events of longer duration or rise times, as may be the case with E4300 keV solar electrons [12]. Accelerations of CMEs have been the focus of considerable work since Sheeley et al. [13] described two classes of
CME speed profiles observed in the range 2–30 R , those gradually accelerating to 400–600 km/s and those with nearly uniform speeds typically in excess of 750 km/s. The second class are candidate drivers of interplanetary shocks and hence important for SEP events [2]. Moon et al. [14] found that the second class is more likely associated with flares and shows decelerations slightly increasing with speeds. Kahler et al. [15] showed that rising time and duration time of SPEs are well correlated with the CME speed; it is consistent result with the shock model of the SPE production. Recently, a new geoeffective CME parameter, which is an earthward direction parameter representing the degree of symmetry of the CME front was suggested and its importance was demonstrated [14,19]. In this study, we examine the correlation between the CME direction parameter and the SPE flux. For this, we use most of the SPEs measured by NOAA’s GOES satellites during solar cycle 23.
2. Data We used SPEs (intensity 410 pfu, with 1 pfu = 1 proton/cm2/s/sr) in the 410 MeV channel of the GOES instrument from 1997 to 2006. GOES data are available on line from NOAA Space Environment Center [16]. Of total 94 SPEs during this period, we studied 63 SPEs whose CME parameters are available. For an estimation of the CME direction parameter, we follow the method suggested by Moon et al. [14] using almost halo CMEs. Then our SPE dataset cover the 67% (63/94) of all the SPEs during solar cycle 23. We investigated LASCO CME catalog based on SOHO coronagraph [17] and flare list from NGDC website based on GOES SXI [18] to examine what physical parameters of solar eruptions might control the SPE intensity and time profile. Fig. 1 shows the schematic diagram of typical gradual SPE time profile. The threshold of SPE (horizontal dashed line) indicates 10 pfu (particle flux unit) threshold at 10 MeV proton energy channel from GOES satellite. The SPE start time is the time of initially facing the threshold line, peak time is when the SPE intensity is on the peak
Fig. 1. Schematic diagram of a gradual SPE time profile.
ARTICLE IN PRESS J. Hwang et al. / Acta Astronautica 67 (2010) 353–361
and the end time is when just leaving the threshold. Rise time (Trise) is the difference between start time and peak time, decrease time (Tdecrease) is the time between peak time and end time, and duration time (Tduration) is the difference between start and end time. Table 1 shows the physical parameters used in this study for three types of events (CME, flare, and SPE). To statistically examine the relationship between the SPE and their associated solar activities, we consider (1) speed, acceleration, angular width, location and earthward direction parameter of each CME, (2) rise time (Trise), duration time (Tduration), decrease time (Tdecrease) and peak intensity (ISPE) of each SPE, and (3) flare strength (x-ray intensity). Fig. 2 describes how to estimate the direction parameter from LASCO C2 running difference images of the 2000 July 14 event. Let us consider the shape of the halo CME, as shown in Fig. 2. If a CME front is directly propagating toward the Earth, the shape in its pre-event subtracted image should be nearly symmetric (like a circle) as shown in Fig. 2. If the CME front is propagating away from the Sun–Earth line, its shape should be quite asymmetric. To quantify its symmetric characteristics, Moon et al. [19] suggest a quantitative parameter as follows: (1) a preevent image is subtracted, (2) an ellipse is plotted on the image and then its major and minor axes are manually adjusted in such a way that the ellipse can approximately Table 1 Associated parameters between SPE, CME and flare. SPE
CME
Flare
Trise Tduration Tdecrease Ispe
Vcme Acceleration Direction parameter
X-ray intensity Solar source longitude
Fig. 2. Example of an estimation of the CME earthward direction parameter on 14 July, 2000.
355
follow the front edge of a CME, (3) straight lines connecting pairs of opposite positions on the CME front are considered, (4) the ratio (b/a) between the shorter distance (b) from the solar disk center and the longer distance (a) is obtained, and (5) its maximum value is finally estimated as the direction parameter; equivalently, the line having the maximum ratio corresponds to an extension of the line connecting solar center and the center of the ellipse. Geometrically, the proposed parameter depends on the ratio of the distance between the ellipse center and solar center to the distance between the center of the ellipse and the CME front. The direction parameter (b/a) of the 14 July, 2000 event of Fig. 2 is 0.64. The closer this value is to the unity, the more the CME is oriented toward the Earth. As a result, the CME of Fig. 2 was associated with a very strong geomagnetic storm (Dst= 300 nT).
3. Analysis and results As mentioned in Section 1, Stolpovskii et al. [12] reported that faster CMEs may drive shocks for longer periods resulting in SEP events of longer duration or rise time. So we can anticipate that some characteristics of the SPEs observed at 1 AU are determined by the properties of the associated CMEs. In order to assess the influence of the CMEs on SPE, we examined the correlations between the various CME properties and the SPE properties. Fig. 3 shows that the relationship between CME speeds and SPE time scales. We found that the CME speed and the SPE duration time are well correlated each other as expected, their linear correlation coefficient is 0.48 and standard deviation is 29.99 (Fig. 3(a)). This result is nearly the same with Kahler et al.’s [15] result although they used totally different spacecraft and period’s dataset. Actually, they compiled a list of E= 20 MeV SEP events observed from 1998 through 2002 with the Energetic Particles: Acceleration, Composition and Transport (EPACT) experiment on the Wind spacecraft. Such a correlation between the CME speed and the SPE total duration time could be explained by the fact that highspeed CMEs can sustain their speeds and become IP shocks during their propagation from the Sun to the Earth, so that they can accelerate particles continuously and produce strong SPEs. It is interesting that the correlation coefficient between SPE decrease time and CME speed is also 0.48, the same as the correlation coefficient between SPE duration time and CME speed (Fig. 3(b)). The correlation coefficient (Fig. 3(c)) between the SPE rise time and the CME speed is 0.34 lower than above two correlation coefficients. This fact indicates the CME speed is well associated with the SPE total duration and decrease time rather than the SPE rise time. The relationship between the CME speed and the SPE peak intensity is shown in Fig. 3(d). The correlation coefficient is 0.31, which is poorer than the other correlations between the SPE time scales and the CME speed. Based on the above results, we can think that (1) the CME speed is better associated with the decline phase of a SPE than its rise phase, and (2) the CME speed might affect to the SPE times more than the SPE intensity.
ARTICLE IN PRESS 356
J. Hwang et al. / Acta Astronautica 67 (2010) 353–361
Fig. 3. Correlation coefficients between the CME speed and (a) the SPE duration time, (b) SPE decrease time, (c) SPE rise time, and (d) SPE peak intensity.
Table 2 Correlation coefficients between SPE, CME and flare. Trise Vcme Direction parameter Acceleration X-ray intensity
0.34 0.2 0.02 0.2
Tduration
Tdecrease
0.48 0.11 0.03 0.32
0.48 0.06 0.05 0.33
Ispe 0.31 0.40 0.08 0.37
We summarized the results of correlation coefficients between the CME/flare parameters, and the SPE parameters in Table 2. The parameters in columns are the SPE times and the SPE peak intensity, and the parameters in rows are associated solar eruption properties such as the CME speed (Vcme), the CME direction parameter, the CME acceleration and the solar flare strength (x-ray intensity). The CME speed shows the best correlation with the SPE duration and the decrease times (cc=0.48 at both cases) as shown in Fig. 3. Of all those the SPE time scales, duration time is the best correlated with the CME speed, which means that the faster the CMEs, the longer the SPEs. The correlation between the CME speed and the
SPE peak intensity is weaker (cc=0.31) than that of the CME speed and the SPE duration time (cc=0.48). So, it seems that the CME speed is more related with the SPE duration time than the SPE peak intensity. It is notable that the CME direction parameter (cc=0.40) is better correlated with the SPE peak intensity than the CME speed (cc=0.31). It indicates that the CME earthward direction might affect to make the SPE intensity higher as the CME speed does. The correlation coefficients between the CME direction parameter and the SPE times are weaker but still positive. It is interesting that the correlations between the CME initial acceleration and the SPE times or intensity are very poor. We note that there are also poor correlations between the CME width and the SPE characteristics, which is not shown here. Such a poor correlation seems to be due to that we chose almost halo CMEs whose direction parameters are available. Actually, of all the dataset, the CME angular widths are larger than 1301. Regarding the solar flares, we compare the flare strength (x-ray intensity) with the SPE parameters. X-ray intensity shows a positive correlation with the SPE peak intensity, but not so noticeable (cc=0.37). This result is somewhat
ARTICLE IN PRESS J. Hwang et al. / Acta Astronautica 67 (2010) 353–361
357
Fig. 4. Relationship between SPE peak intensity and CME direction parameter.
Fig. 5. Histogram of the SPE peak flux.
disappointing when considering other previous reports on x-ray strength affecting the SPE peak intensity. In this study, we found a new CME parameter, the direction parameter, which is related to the SPE peak intensity. The CME direction parameter shows a correlation with the SPE peak intensity (cc=0.40), even comparable to x-ray intensity with the SPE peak intensity (cc=0.37). As shown in Table 2, it is clear that the CME speed is correlated with the SPE time scales (rise, duration, and decrease time), that is, there is a tendency that CME speed is related to the SPE time scales. On the other hands, the CME direction parameter together with x-ray strength makes the SPE intensity higher. While investigating the correlation between the CME direction parameter and the SPE peak intensity, we found a very interesting result. Fig. 4 shows there exists a clear separation between two groups; one group shows no correlation between the CME direction parameter and the SPE peak intensity, shows almost a linear correlation. The correlation coefficient between the CME direction parameter and the SPE peak intensity for all 63 dataset is just 0.38, but for extremely large six SPEs (proton peak intensity is over 104 pfu), the correlation coefficient between them is 0.65 as shown at Fig. 4. It is evidently in contrast with no correlation (cc=0.01) for remaining 57 dataset. The more detailed inspection on what makes this clear discrimination between two groups will be studied in the following section. In Fig. 4, the red circle indicates the solar source longitude is central; east 301–west 301, a blue circle indicates west 30–901; a the green circle indicates east 30–901. It is noted that the source locations of those extremely large six SPEs are almost central except only one case (8 November, 2000). If we include the event the flux of which is near 5,000 pfu, there are seven events standing almost in a line.
the peak flux of all the other 57 events are less than 10,000 pfu. If we include one event over 5,000 pfu, it seems that there are total seven SPEs which show a similar dependence to the CME direction parameter. Actually, the large six events cover the largest six SPEs during the last solar cycle 23. The CME and solar eruption information of those large six events are given in Table 3. Fig. 6 shows the distribution of all 63 events: (a) the proton flux, (b) the direction parameter, (c) the CME speed and (d) the x-ray intensity. Gray-colored circles indicate the large six events. The size of a circle indicates a size of each parameter. Fig. 6(a) shows that most of large six events are located near the center except for only one event (8 November, 2000) which is located near the western limb (N00-10, W75-80). Fig. 6(b) shows the distribution of the direction parameter. It is interesting that all SPEs are located between N22 and S16 and there is a similar tendency between (a) and (b) showing proportional relationship of the large six events. Figs. 6(c) and (d) show the CME speed and the x-ray flux distribution each, for two parameters, which do not have any dependence on the SPE peak intensity. It is known that intensity–time profiles of the SPEs depend on the longitude of the solar event relative to the observer. Cane et al. [7] suggested that change in shape as a function of heliolongitude within the framework of a recently derived model for the large-scale structure of IP shocks. In particular, a long delay to the maximum intensity for the far eastern events and the overall extended duration can be accounted for by IP shock acceleration and a continued magnetic connection to the shock even after it has propagated beyond 1 AU. In this study, we found not only the time profile of the SPE but also the SPE intensities themselves are affected by the longitude of the solar events. Therefore, our large six SPEs located at disk center are well-connected to the shock and to the earth, and show the good correlation to the CME earthward direction parameter. Regarding the origins of the SPEs, there is still no widely accepted theory that explains all the observed properties of the SPEs. The seed particles may originate
4. What makes this good correlation between the CME direction parameter and the SPE peak intensity? Fig. 5 shows the histogram of the SPE peak intensity distribution of all 63 SPEs. Excluding the large six SPEs,
ARTICLE IN PRESS 358
J. Hwang et al. / Acta Astronautica 67 (2010) 353–361
Table 3 Large six events’ characteristics. No.
CME date/UT
CME speed
CME direction parameter
Location
AR
Flare strength
Flare class
Proton flux
1 2 3 4 5 6
14/07/2000; 08/11/2000; 24/09/2001; 04/11/2001; 22/11/2001; 28/10/2003;
1674 1738 2402 1810 1437 2459
0.75 0.45 0.37 0.77 0.60 0.94
N22W07 N00-10W75-80 S16E23 N06W18 S15W34 S16E08
9077 9690 9632 9684 9704 10486
0.75 0.21 0.63 0.22 0.31 1.8
X05/3B M74/multiple X02/2B X01/3B M09/2N X17/4B
24000 14800 12900 31700 18900 29500
10:54 23:06 10:30 16:35 23:30 11:30
N
N
E
W
W
E
S
S
N
N
E
W
S
W
E
S
Fig. 6. Distribution of Locations of the CMEs associated with the SPE, the size of circles indicates (a) a proton peak intensity, (b) a CME direction parameter, (c) a CME speed and (d) x-ray intensity. Six closed circles indicate the large six events.
from impulsive flares [20] or from preceding CMEs [21]. In addition to the presence of seed particles, the physical conditions in the ambient medium may also modify the characteristics of shocks and hence affect the intensity of the SPEs [22]. In order to assess the influence of interplanetary medium, we searched for other CMEs from the same source region as the primary CME that satisfies the following criteria: (1) the preceding CMEs should have been launched within 24 hours ahead of the onset time of the primary CMEs; (2) its angular width is wider than 601; and (3) both CMEs were ejected from a similar source region and a similar ejected direction. We adopted the angular width condition from Gopalswamy et al.’s [22] criterion which requires CME width must be wider than average CME width of non-halo, 471. We found
that all the large six events have preceding CMEs satisfying above criteria as shown in Fig. 7. Actually, four preceding CMEs of six events are already reported by Gopalswamy et al. [23]. We found two more events, since they examined the dataset between 1997 and 2001 and we investigated the period of 1997–2006. Fig. 8 shows that there are also streamers which interact with the following primary CMEs. The disturbance in the streamer structure after a CME rejected is one of the evidence of the interaction between a streamer and a CME. Another evidence of a streamer–CME interaction is the existence of a type II radio burst. Since the type II radio burst might be produced at the CME flanks which interacts with a nearby high-density streamer (low Alfvenic region) even
ARTICLE IN PRESS J. Hwang et al. / Acta Astronautica 67 (2010) 353–361
359
2001/09/24
2000/07/14
Halo (S)
CPA :167°, angular width: 163
°
CPA: 271°, angular width: 61°
Halo (OA)
2001/11/04
2001/11/22
Halo (BA) Halo (BA)
Halo (BA)
Halo (OA) 2003/10/28
2000/11/08
Halo (BA)
CPA: 124°, angular width: 147°
CPA: 271°, angular width: >170°
Halo (S)
Fig. 7. Preceding CMEs before the primary CMEs for the large six events.
though the CME is not so fast [24]. We found there are type II radio burst in all six events. According to Gopalswamy et al. [22], the SPE intensities are similar between CME–CME and CME–streamer interaction groups. Therefore, it seems that the effect of a streamer on the SPE is similar to that of a preceding CME on the SPE. As for the role of the direction parameter, it is interesting to note that the direction parameter is intimately related to the SPE flux, which may be considered to predict a large SPE arriving at 1 AU. If a solar eruption occurs at disk center, the prediction efficiency might be higher. The value of the CME earthward direction parameter as an important geoeffectiveness indicator has been established by other previous researchers [19,25]. We suggest that the direction parameter could be considered as an important input parameter for empirically predicting large SPEs. 5. Summary and discussion We have examined the solar proton events (SPEs) associated with the coronal mass ejections (CMEs) during
the solar cycle 23 (1997–2006) to understand the correlations between the SPE characteristics and the CME parameters. While the overall correlation is similar to what was found before [2,23], we found a new result between the CME direction parameter and the SPE intensity. Using total 63 SPEs, we investigated the relationships among SPEs, CMEs and flares, and found that (1) the SPE rise, duration and decrease time depend on the CME speed, and (2) the SPE peak intensity depends on the CME earthward direction parameter and x-ray flare intensity. As expected, we found relatively good correlations between the CME speed and the SPE time scales such as rise, duration, and decrease times, which mean that the faster CMEs may drive shocks for a longer period, resulting in the SPEs of longer duration and decrease times. While inspecting the relation between the SPE peak intensity and the direction parameter, we found there are two groups: the first group consists of extremely large six SPEs (proton peak intensity is over 104 pfu); the correlation coefficients is 0.65, which is the highest correlation in this correlation study and there is no one
ARTICLE IN PRESS 360
J. Hwang et al. / Acta Astronautica 67 (2010) 353–361
2001/09/24
2000/07/14
2001/11/22
2001/11/04
2003/10/28
2000/11/08
Fig. 8. Streamer–CME interactions for the large six events.
to find such a high correlation in many previous statistical studies. Such a high correlation coefficient is in contrast with cc=0.01 for the remaining 57 dataset excluding large six events. This clear separation might be caused by different acceleration mechanisms or different acceleration efficiency, so we do investigate the characteristics and solar environments of the large six events in detail. By investigating characteristics of the first group, we found that all the SPEs are associated with very fast halo CME ( 41400km/s). We also found that these large six events were (except for one event) located at disk center within a latitude strip between N22 and S16, and a longitude band between W34 and E23. According to Cane et al. [7], a shape of the SPE time profile is a function of heliolongitude within the framework of a recently derived model for the large-scale structure of IP shocks. In this study, we found the SPE intensities themselves are affected by not only the longitude of the solar event but also its time profile. Therefore, the large six SPEs located at disk center are well-connected to the shock and to the earth, and show the good correlation to the CME earthward direction parameter. In other words, the CME
earthward direction parameter can be an important input parameter of an empirical prediction model for the SPEs arriving at 1 AU, especially if the source longitude is located at disk center and the associated CME is very fast. In this study, we found a strong correlation between the SPE peak intensity and the CME earthward direction parameter for the large six SPEs, while the correlations between all other parameters were found to be marginal or poor. The reasons why the SPE intensity is proportional to the CME earthward direction parameter could be understood by considering the following facts: (1) if we apply the projection effects depending on direction parameter to the six CMEs, a CME with a large direction parameter will gain faster speeds and become an extremely fast CMEs. (2) We may suppose that, there is an IMF connection from a CME front to the earth for the CME with the large direction parameter and that from a CME flank to the earth for the CME with the small direction parameter. The shock ahead of the CME front with the large direction parameter should be stronger than that from the CME flank so that the shock could accelerate particles more efficiently than that from the
ARTICLE IN PRESS J. Hwang et al. / Acta Astronautica 67 (2010) 353–361
CME flank with the small direction parameter. In addition, the particles from the CME front with the large directional parameter could propagate directly along the IMF line connecting the CME front and the earth. (3) Extremely fast CMEs with large direction parameters near the Sun will sustain their speeds and become strong IP shocks during the Sun to the earth. The strong IP shock can accelerate particles continuously and produce strong SPEs. We found that there is the good correlation between the SPE peak intensity and the CME earthward direction parameter only for the large six events while there is no correlation for the low-intensity SPEs. One may have a question what makes a clear separation between those two groups. To answer this question we examined the large six SPEs in detail and noted that there are preceding CMEs originated from the same solar source region, and there exists a nearby helmet streamer. We speculate that the preceding CME and pre-existing helmet streamer might provide seed particles by the following CME-driven shocks. Our results support the requirement for the highflux SPE, which was proposed by Gopalswamy et al. [22], i.e., the primary fast CME is originated from the same active region as that of the preceding CME and interacts with the preceding slow CME. The presence of the preexisting streamer might give the similar effect with the preceding CME. For SPEs in which the primary CME interacted with the streamer or with the preceding CME very close in time, proton intensities were highly correlated with the CME direction parameter. In this respect, a particle acceleration mechanism of SPEs associated with the CME–CME interaction and the CME–streamer interaction needs to be investigated in the near future.
Acknowledgement This work has been supported by the ‘Development of Korean Space Weather Center’ of KASI and the KASI basic research funds. Data of space satellite of NGDC, CDAW and GOES have been utilized. Yong-Jae Moon has been supported by the WCU grant (No. R31-10016) funded by the Korean Ministry of Education, Science and Technology
361
and by the Korea Research Foundation Grant funded by the Korean Government (MOEHRD, Basic Research Promotion Fund) (KRF-2008-314-C00158, 20090071744). References [1] D.V. Reames, Space Sci. Rev 90 (1999) 413. [2] S.W. Kahler, J. Geophys. Res. 106 (2001) 20, 947. [3] N. Gopalswamy, S. Yashiro, M.L. Kaiser, R.A. Howard, J.-L. Bougeret, Astrophys. J. 548 (2001) L91. [4] N. Gopalswamy, S. Yashiro, M.L. Kaiser, R.A. Howard, J.-L. Bougeret, Geophys. Res. Lett. 29 (8) (2002) 1265, doi:10.1029/2001GL013606. [5] N. Gopalswamy, S. Yashiro, A. Lara, M.L. Kaiser, B.J. Thompson, P.T. Gallagher, R.A. Howard, Geophys. Res. Lett. 30 (12) (2003) 8015, doi:10.1029/2002GL016435. [6] M.A.I. van Hollebeke, L.S. Ma Sung, F.B. McDonald, Sol. Phys. 41 (1975) 189. [7] H.V. Cane, D.V. Reames, T.T. von Rosenvinge, J. Geophys. Res. 93 (1988) 9555. [8] C.C. Balch, Radiat. Meas. 30 (1999) 231. [9] D.V. Reames, L.M. Barbier, C.K. Ng, Astrophys. J. 466 (1996) 473. [10] D.V. Reames, S.W. Kahler, C.K. Ng, Astrophys. J. 491 (1997) 414. [11] E.I. Daibog, Yu.I. Logachev, S. Kahler, K. Kecskemety, Adv. Space Res. 32 (12) (2003) 2655. [12] V.G. Stolpovskii, E.I. Daibog, S.W. Kahler, G. Erdos, Adv. Space Res. 21 (1998) 543. [13] N.R. Sheeley Jr., J.H. Walters, Y.-M. Wang, R.A. Howard, J. Geophys. Res. 104 (1999) 24739. [14] Y.-J. Moon, G.S. Choe, H. Wang, Y.D. Park, N. Gopalswamy, G. Yang, S. Yashiro, Astrophys. J. 581 (2002) 694. [15] S.W. Kahler, Astrophys. J 628 (2005) 1014. [16] /http://sec.noaa.govS. [17] /http://cdaw.gsfc.nasa.gov/CME_list/UNIVERSAL/2003_07/ univ2003_07.htmlS. [18] /http://sxi.ngdc.noaa.govS. [19] Y.-J. Moon, K.-S. Cho, M. Dryer, Y.-H. Kim, Su-chan Bong, Jongchul Chae, Y.D. Park, Astrophys. J. 624 (2005) 414. [20] G.M. Mason, J.E. Mazur, J.R. Dwyer, 3HE enhancements in large solar energetic particle events, Astrophys. J. 525 (1999) L133. [21] N. Gopalswamy, S. Yashiro, G. Michaek, M.L. Kaiser, R.A. Howard, D.V. Reames, R. Leske, T. von Rosenvinge, Astrophys. J. 572 (2002) L103–L107. [22] N. Gopalswamy, S. Yashiro, S. Krucker, G. Stenborg, R.A. Howard, J. Geophys. Res. 109 (2004) A12105, doi:10.1029/2004JA010602. [23] N. Gopalswamy, S. Yashiro, G. Michalek, M. L. Kaiser, R. A. Howard, R. Leske, T. von Rosenvinge, D. V. Reames, in: M. Velli, R. Bruno, F. Malara (Eds.), AIP Conf. Proc., 2003, 679, 608. [24] K.-S. Cho, S.-C. Bong, Y.-H. Kim, Y.-J. Moon, M. Dryer, A. Shanmugaraju, J. Lee, Y.D. Park, A&A 491 (2008) 873, doi:10.1051/0004-6361:20079013. [25] R.-S. Kim, K.-S. Cho, K.-H. Kim, Y.-D. Park, Y.-J. Moon, Y. Yi, J. Lee, H. Wang, H. Song, M. Dryer, Astrophys. J. 677 (2008) 1378–1384.
Contents lists available at ScienceDirect
Acta Astronautica journal homepage: www.elsevier.com/locate/actaastro
Solar proton events during the solar cycle 23 and their association with CME parameters$ Junga Hwang a,, Kyung-Suk Cho a, Young-Jae Moon b, Rok-Soon Kim c, Young-Deuk Park a a b c
Korea Astronomy and Space Science Institute, 61-1, Hwaam dong, Yuseong gu Daejeon, 305-348, South Korea School of Space Research, Kyung Hee University, Yongin-si, Korea Chungnam National University, Daejeon, Korea
a r t i c l e in fo
abstract
Article history: Received 22 January 2010 Received in revised form 11 March 2010 Accepted 5 April 2010 Available online 28 April 2010
We have studied the solar proton events associated with the coronal mass ejections (CMEs) and flares during the solar cycle 23 (1997–2006) in order to determine what physical parameters of the solar eruptions might control the SPE intensity and time profile. For total 63 SPEs, we found that (1) SPE rise time, duration time and decrease times depend on a CME speed (cc= 0.34, 0.48 and 0.48) and (2) a SPE peak intensity depends on an earthward direction parameter of a CME as well as the CME speed and xray flare intensity (cc =0.40, 0.31 and 0.37). The SPEs were divided into two groups according to the correlation between the CME earthward direction parameter and the SPE intensity. First group consists of large six SPEs (>10,000 pfu at >10 MeV proton channel of GOES satellite) and shows a very good correlation (cc= 0.65) between the SPE peak intensity and the CME earthward direction parameter. Second group has a relatively weak SPE peak intensity and shows no correlation (cc= 0.01) between the SPE peak intensity and the CME earthward direction parameter we found that the first group SPEs are associated with a very fast halo CME (>1400 km/s) and most of those are located at disk except for only one case. Especially, large six SPEs have a good correlation with their associated CME earthward direction parameters, implying that these events are produced by ICME-driven shocks. We also found that those six SPEs are associated with the preceding CMEs originated from the same solar source region and a nearby pre-existing helmet streamer. Thus, we speculate that the preceding CME and helmet streamer might provide seed particles for CME-driven shocks and cause a clear separation between two groups. & 2010 Elsevier Ltd. All rights reserved.
Keywords: Solar energetic particle Corona mass ejection Solar flare
1. Introduction Solar flares and coronal mass ejections (CMEs) are the most powerful events in the solar system. In tens of minutes they can convert in excess of 1032 ergs of magnetic energy into accelerated particles, heated plasma, and ejected solar
$
This paper was presented during the 60th IAC in Daejeon.
Corresponding author. Tel.: + 82 42 865 2061; fax: + 82 42 865 2020.
E-mail addresses: [email protected], [email protected] (J. Hwang), [email protected] (K.-S. Cho), [email protected] (Y.-J. Moon), [email protected] (R.-S. Kim), [email protected] (Y.-D. Park). 0094-5765/$ - see front matter & 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.actaastro.2010.04.001
material. In this paper, we chose the word of ‘‘Solar Proton Events (SPEs)’’ instead of ‘‘Solar Energetic Particles (SEPs)’’ as we analyzed only proton data. Some large SPEs had a rise time (from a threshold to the maximum intensity) within tens of minutes. This rapid rise time, coupled with the high intensity, means that if astronauts had been on EVA (Extra Vehicular Activity) on the Moon, there would be very little warning before the significant enhancement of the SPE flux. Events of this nature need to be taken into account in planning astronaut operations and shelters on the Moon and in interplanetary space, which is because the space radiation exposure to a human body and spacecraft can make severe
ARTICLE IN PRESS 354
J. Hwang et al. / Acta Astronautica 67 (2010) 353–361
radiation hazards. In addition, the rapid rise places extreme requirements on models that accelerate SPEs by CME-driven shocks, because protons with energies 41 GeV must have been accelerated within minutes in the very low corona. There is a substantial evidence that gradual SPEs observed at 1 AU result from an acceleration at coronal and interplanetary shocks driven by fast CMEs [1,2]. CME speeds have been found to be correlated with the peak intensities of the associated SPEs, but there is a very broad scatter. This weak correlation indicates a complexity in the SPE shock acceleration process that has not been fully understood. Recently, several attempts have been made to explore the effects that variation in the coronal magnetic and particle environment might have in the resulting SPE production by CME-driven shocks. Enhanced decametric– hectometric (DH) emission was observed during interactions of a fast CME with a preceding slower CME [3,4]. The enhanced DH emission was interpreted as evidence for strengthening of a pre-existing shock of the slower CME and for the formation of a new shock in the fast CME. The basic interpretation was that the efficiency of particle acceleration is somehow enhanced when the primary CME runs into regions of enhanced density due to preceding CMEs or streamers [5]. The first large statistical study of E420 MeV SPEs was probably that of van Hollebeke et al. [6]. Using 185 20 MeV oEo80 MeV SPEs from the Interplanetary Monitoring Platform (IMP) spacecraft, they focused on the early phases of SEP events and plotted times DT, from event onsets to maxima, as a function of solar source longitude. The time DT reached a minimum value of about 2.5 h around W501 and increased for larger magnetic angular separations between the Earth and the solar source longitude. Cane et al. [7] surveyed a larger sample of 235 SEP events and plotted as a function of solar source longitude the delay times from associated Ha flare maxima to peak SEP intensities, some of which were peaks at 1 AU shock passages. In each of three proton energy ranges spanning 1–100 MeV, those plots showed a pattern of delay times increasing from W901 to E901 source longitudes. A similar result was found for 112 E4 10 MeV proton events by Balch [8]. The decay phases of gradual SEP events are characterized by a spatial and temporal invariance of the energy spectra, whose onset is ordered by the location of the observer relative to the interplanetary shock [9,10]. Statistical analyses of E4 4 MeV proton event decay phases show a tendency for event decay times to decrease with higher shock speeds and with steeper energy spectra [11], supporting the interpretation of adiabatic deceleration of SEPs quasitrapped behind large-scale expanding shocks. Since CMEs are the drivers of shocks that accelerate SEPs, we might expect that the characteristics of the SEP intensity–time profiles observed at 1 AU are determined by properties of the associated CMEs. For example, faster CMEs may drive shocks for longer periods of time, resulting in SEP events of longer duration or rise times, as may be the case with E4300 keV solar electrons [12]. Accelerations of CMEs have been the focus of considerable work since Sheeley et al. [13] described two classes of
CME speed profiles observed in the range 2–30 R , those gradually accelerating to 400–600 km/s and those with nearly uniform speeds typically in excess of 750 km/s. The second class are candidate drivers of interplanetary shocks and hence important for SEP events [2]. Moon et al. [14] found that the second class is more likely associated with flares and shows decelerations slightly increasing with speeds. Kahler et al. [15] showed that rising time and duration time of SPEs are well correlated with the CME speed; it is consistent result with the shock model of the SPE production. Recently, a new geoeffective CME parameter, which is an earthward direction parameter representing the degree of symmetry of the CME front was suggested and its importance was demonstrated [14,19]. In this study, we examine the correlation between the CME direction parameter and the SPE flux. For this, we use most of the SPEs measured by NOAA’s GOES satellites during solar cycle 23.
2. Data We used SPEs (intensity 410 pfu, with 1 pfu = 1 proton/cm2/s/sr) in the 410 MeV channel of the GOES instrument from 1997 to 2006. GOES data are available on line from NOAA Space Environment Center [16]. Of total 94 SPEs during this period, we studied 63 SPEs whose CME parameters are available. For an estimation of the CME direction parameter, we follow the method suggested by Moon et al. [14] using almost halo CMEs. Then our SPE dataset cover the 67% (63/94) of all the SPEs during solar cycle 23. We investigated LASCO CME catalog based on SOHO coronagraph [17] and flare list from NGDC website based on GOES SXI [18] to examine what physical parameters of solar eruptions might control the SPE intensity and time profile. Fig. 1 shows the schematic diagram of typical gradual SPE time profile. The threshold of SPE (horizontal dashed line) indicates 10 pfu (particle flux unit) threshold at 10 MeV proton energy channel from GOES satellite. The SPE start time is the time of initially facing the threshold line, peak time is when the SPE intensity is on the peak
Fig. 1. Schematic diagram of a gradual SPE time profile.
ARTICLE IN PRESS J. Hwang et al. / Acta Astronautica 67 (2010) 353–361
and the end time is when just leaving the threshold. Rise time (Trise) is the difference between start time and peak time, decrease time (Tdecrease) is the time between peak time and end time, and duration time (Tduration) is the difference between start and end time. Table 1 shows the physical parameters used in this study for three types of events (CME, flare, and SPE). To statistically examine the relationship between the SPE and their associated solar activities, we consider (1) speed, acceleration, angular width, location and earthward direction parameter of each CME, (2) rise time (Trise), duration time (Tduration), decrease time (Tdecrease) and peak intensity (ISPE) of each SPE, and (3) flare strength (x-ray intensity). Fig. 2 describes how to estimate the direction parameter from LASCO C2 running difference images of the 2000 July 14 event. Let us consider the shape of the halo CME, as shown in Fig. 2. If a CME front is directly propagating toward the Earth, the shape in its pre-event subtracted image should be nearly symmetric (like a circle) as shown in Fig. 2. If the CME front is propagating away from the Sun–Earth line, its shape should be quite asymmetric. To quantify its symmetric characteristics, Moon et al. [19] suggest a quantitative parameter as follows: (1) a preevent image is subtracted, (2) an ellipse is plotted on the image and then its major and minor axes are manually adjusted in such a way that the ellipse can approximately Table 1 Associated parameters between SPE, CME and flare. SPE
CME
Flare
Trise Tduration Tdecrease Ispe
Vcme Acceleration Direction parameter
X-ray intensity Solar source longitude
Fig. 2. Example of an estimation of the CME earthward direction parameter on 14 July, 2000.
355
follow the front edge of a CME, (3) straight lines connecting pairs of opposite positions on the CME front are considered, (4) the ratio (b/a) between the shorter distance (b) from the solar disk center and the longer distance (a) is obtained, and (5) its maximum value is finally estimated as the direction parameter; equivalently, the line having the maximum ratio corresponds to an extension of the line connecting solar center and the center of the ellipse. Geometrically, the proposed parameter depends on the ratio of the distance between the ellipse center and solar center to the distance between the center of the ellipse and the CME front. The direction parameter (b/a) of the 14 July, 2000 event of Fig. 2 is 0.64. The closer this value is to the unity, the more the CME is oriented toward the Earth. As a result, the CME of Fig. 2 was associated with a very strong geomagnetic storm (Dst= 300 nT).
3. Analysis and results As mentioned in Section 1, Stolpovskii et al. [12] reported that faster CMEs may drive shocks for longer periods resulting in SEP events of longer duration or rise time. So we can anticipate that some characteristics of the SPEs observed at 1 AU are determined by the properties of the associated CMEs. In order to assess the influence of the CMEs on SPE, we examined the correlations between the various CME properties and the SPE properties. Fig. 3 shows that the relationship between CME speeds and SPE time scales. We found that the CME speed and the SPE duration time are well correlated each other as expected, their linear correlation coefficient is 0.48 and standard deviation is 29.99 (Fig. 3(a)). This result is nearly the same with Kahler et al.’s [15] result although they used totally different spacecraft and period’s dataset. Actually, they compiled a list of E= 20 MeV SEP events observed from 1998 through 2002 with the Energetic Particles: Acceleration, Composition and Transport (EPACT) experiment on the Wind spacecraft. Such a correlation between the CME speed and the SPE total duration time could be explained by the fact that highspeed CMEs can sustain their speeds and become IP shocks during their propagation from the Sun to the Earth, so that they can accelerate particles continuously and produce strong SPEs. It is interesting that the correlation coefficient between SPE decrease time and CME speed is also 0.48, the same as the correlation coefficient between SPE duration time and CME speed (Fig. 3(b)). The correlation coefficient (Fig. 3(c)) between the SPE rise time and the CME speed is 0.34 lower than above two correlation coefficients. This fact indicates the CME speed is well associated with the SPE total duration and decrease time rather than the SPE rise time. The relationship between the CME speed and the SPE peak intensity is shown in Fig. 3(d). The correlation coefficient is 0.31, which is poorer than the other correlations between the SPE time scales and the CME speed. Based on the above results, we can think that (1) the CME speed is better associated with the decline phase of a SPE than its rise phase, and (2) the CME speed might affect to the SPE times more than the SPE intensity.
ARTICLE IN PRESS 356
J. Hwang et al. / Acta Astronautica 67 (2010) 353–361
Fig. 3. Correlation coefficients between the CME speed and (a) the SPE duration time, (b) SPE decrease time, (c) SPE rise time, and (d) SPE peak intensity.
Table 2 Correlation coefficients between SPE, CME and flare. Trise Vcme Direction parameter Acceleration X-ray intensity
0.34 0.2 0.02 0.2
Tduration
Tdecrease
0.48 0.11 0.03 0.32
0.48 0.06 0.05 0.33
Ispe 0.31 0.40 0.08 0.37
We summarized the results of correlation coefficients between the CME/flare parameters, and the SPE parameters in Table 2. The parameters in columns are the SPE times and the SPE peak intensity, and the parameters in rows are associated solar eruption properties such as the CME speed (Vcme), the CME direction parameter, the CME acceleration and the solar flare strength (x-ray intensity). The CME speed shows the best correlation with the SPE duration and the decrease times (cc=0.48 at both cases) as shown in Fig. 3. Of all those the SPE time scales, duration time is the best correlated with the CME speed, which means that the faster the CMEs, the longer the SPEs. The correlation between the CME speed and the
SPE peak intensity is weaker (cc=0.31) than that of the CME speed and the SPE duration time (cc=0.48). So, it seems that the CME speed is more related with the SPE duration time than the SPE peak intensity. It is notable that the CME direction parameter (cc=0.40) is better correlated with the SPE peak intensity than the CME speed (cc=0.31). It indicates that the CME earthward direction might affect to make the SPE intensity higher as the CME speed does. The correlation coefficients between the CME direction parameter and the SPE times are weaker but still positive. It is interesting that the correlations between the CME initial acceleration and the SPE times or intensity are very poor. We note that there are also poor correlations between the CME width and the SPE characteristics, which is not shown here. Such a poor correlation seems to be due to that we chose almost halo CMEs whose direction parameters are available. Actually, of all the dataset, the CME angular widths are larger than 1301. Regarding the solar flares, we compare the flare strength (x-ray intensity) with the SPE parameters. X-ray intensity shows a positive correlation with the SPE peak intensity, but not so noticeable (cc=0.37). This result is somewhat
ARTICLE IN PRESS J. Hwang et al. / Acta Astronautica 67 (2010) 353–361
357
Fig. 4. Relationship between SPE peak intensity and CME direction parameter.
Fig. 5. Histogram of the SPE peak flux.
disappointing when considering other previous reports on x-ray strength affecting the SPE peak intensity. In this study, we found a new CME parameter, the direction parameter, which is related to the SPE peak intensity. The CME direction parameter shows a correlation with the SPE peak intensity (cc=0.40), even comparable to x-ray intensity with the SPE peak intensity (cc=0.37). As shown in Table 2, it is clear that the CME speed is correlated with the SPE time scales (rise, duration, and decrease time), that is, there is a tendency that CME speed is related to the SPE time scales. On the other hands, the CME direction parameter together with x-ray strength makes the SPE intensity higher. While investigating the correlation between the CME direction parameter and the SPE peak intensity, we found a very interesting result. Fig. 4 shows there exists a clear separation between two groups; one group shows no correlation between the CME direction parameter and the SPE peak intensity, shows almost a linear correlation. The correlation coefficient between the CME direction parameter and the SPE peak intensity for all 63 dataset is just 0.38, but for extremely large six SPEs (proton peak intensity is over 104 pfu), the correlation coefficient between them is 0.65 as shown at Fig. 4. It is evidently in contrast with no correlation (cc=0.01) for remaining 57 dataset. The more detailed inspection on what makes this clear discrimination between two groups will be studied in the following section. In Fig. 4, the red circle indicates the solar source longitude is central; east 301–west 301, a blue circle indicates west 30–901; a the green circle indicates east 30–901. It is noted that the source locations of those extremely large six SPEs are almost central except only one case (8 November, 2000). If we include the event the flux of which is near 5,000 pfu, there are seven events standing almost in a line.
the peak flux of all the other 57 events are less than 10,000 pfu. If we include one event over 5,000 pfu, it seems that there are total seven SPEs which show a similar dependence to the CME direction parameter. Actually, the large six events cover the largest six SPEs during the last solar cycle 23. The CME and solar eruption information of those large six events are given in Table 3. Fig. 6 shows the distribution of all 63 events: (a) the proton flux, (b) the direction parameter, (c) the CME speed and (d) the x-ray intensity. Gray-colored circles indicate the large six events. The size of a circle indicates a size of each parameter. Fig. 6(a) shows that most of large six events are located near the center except for only one event (8 November, 2000) which is located near the western limb (N00-10, W75-80). Fig. 6(b) shows the distribution of the direction parameter. It is interesting that all SPEs are located between N22 and S16 and there is a similar tendency between (a) and (b) showing proportional relationship of the large six events. Figs. 6(c) and (d) show the CME speed and the x-ray flux distribution each, for two parameters, which do not have any dependence on the SPE peak intensity. It is known that intensity–time profiles of the SPEs depend on the longitude of the solar event relative to the observer. Cane et al. [7] suggested that change in shape as a function of heliolongitude within the framework of a recently derived model for the large-scale structure of IP shocks. In particular, a long delay to the maximum intensity for the far eastern events and the overall extended duration can be accounted for by IP shock acceleration and a continued magnetic connection to the shock even after it has propagated beyond 1 AU. In this study, we found not only the time profile of the SPE but also the SPE intensities themselves are affected by the longitude of the solar events. Therefore, our large six SPEs located at disk center are well-connected to the shock and to the earth, and show the good correlation to the CME earthward direction parameter. Regarding the origins of the SPEs, there is still no widely accepted theory that explains all the observed properties of the SPEs. The seed particles may originate
4. What makes this good correlation between the CME direction parameter and the SPE peak intensity? Fig. 5 shows the histogram of the SPE peak intensity distribution of all 63 SPEs. Excluding the large six SPEs,
ARTICLE IN PRESS 358
J. Hwang et al. / Acta Astronautica 67 (2010) 353–361
Table 3 Large six events’ characteristics. No.
CME date/UT
CME speed
CME direction parameter
Location
AR
Flare strength
Flare class
Proton flux
1 2 3 4 5 6
14/07/2000; 08/11/2000; 24/09/2001; 04/11/2001; 22/11/2001; 28/10/2003;
1674 1738 2402 1810 1437 2459
0.75 0.45 0.37 0.77 0.60 0.94
N22W07 N00-10W75-80 S16E23 N06W18 S15W34 S16E08
9077 9690 9632 9684 9704 10486
0.75 0.21 0.63 0.22 0.31 1.8
X05/3B M74/multiple X02/2B X01/3B M09/2N X17/4B
24000 14800 12900 31700 18900 29500
10:54 23:06 10:30 16:35 23:30 11:30
N
N
E
W
W
E
S
S
N
N
E
W
S
W
E
S
Fig. 6. Distribution of Locations of the CMEs associated with the SPE, the size of circles indicates (a) a proton peak intensity, (b) a CME direction parameter, (c) a CME speed and (d) x-ray intensity. Six closed circles indicate the large six events.
from impulsive flares [20] or from preceding CMEs [21]. In addition to the presence of seed particles, the physical conditions in the ambient medium may also modify the characteristics of shocks and hence affect the intensity of the SPEs [22]. In order to assess the influence of interplanetary medium, we searched for other CMEs from the same source region as the primary CME that satisfies the following criteria: (1) the preceding CMEs should have been launched within 24 hours ahead of the onset time of the primary CMEs; (2) its angular width is wider than 601; and (3) both CMEs were ejected from a similar source region and a similar ejected direction. We adopted the angular width condition from Gopalswamy et al.’s [22] criterion which requires CME width must be wider than average CME width of non-halo, 471. We found
that all the large six events have preceding CMEs satisfying above criteria as shown in Fig. 7. Actually, four preceding CMEs of six events are already reported by Gopalswamy et al. [23]. We found two more events, since they examined the dataset between 1997 and 2001 and we investigated the period of 1997–2006. Fig. 8 shows that there are also streamers which interact with the following primary CMEs. The disturbance in the streamer structure after a CME rejected is one of the evidence of the interaction between a streamer and a CME. Another evidence of a streamer–CME interaction is the existence of a type II radio burst. Since the type II radio burst might be produced at the CME flanks which interacts with a nearby high-density streamer (low Alfvenic region) even
ARTICLE IN PRESS J. Hwang et al. / Acta Astronautica 67 (2010) 353–361
359
2001/09/24
2000/07/14
Halo (S)
CPA :167°, angular width: 163
°
CPA: 271°, angular width: 61°
Halo (OA)
2001/11/04
2001/11/22
Halo (BA) Halo (BA)
Halo (BA)
Halo (OA) 2003/10/28
2000/11/08
Halo (BA)
CPA: 124°, angular width: 147°
CPA: 271°, angular width: >170°
Halo (S)
Fig. 7. Preceding CMEs before the primary CMEs for the large six events.
though the CME is not so fast [24]. We found there are type II radio burst in all six events. According to Gopalswamy et al. [22], the SPE intensities are similar between CME–CME and CME–streamer interaction groups. Therefore, it seems that the effect of a streamer on the SPE is similar to that of a preceding CME on the SPE. As for the role of the direction parameter, it is interesting to note that the direction parameter is intimately related to the SPE flux, which may be considered to predict a large SPE arriving at 1 AU. If a solar eruption occurs at disk center, the prediction efficiency might be higher. The value of the CME earthward direction parameter as an important geoeffectiveness indicator has been established by other previous researchers [19,25]. We suggest that the direction parameter could be considered as an important input parameter for empirically predicting large SPEs. 5. Summary and discussion We have examined the solar proton events (SPEs) associated with the coronal mass ejections (CMEs) during
the solar cycle 23 (1997–2006) to understand the correlations between the SPE characteristics and the CME parameters. While the overall correlation is similar to what was found before [2,23], we found a new result between the CME direction parameter and the SPE intensity. Using total 63 SPEs, we investigated the relationships among SPEs, CMEs and flares, and found that (1) the SPE rise, duration and decrease time depend on the CME speed, and (2) the SPE peak intensity depends on the CME earthward direction parameter and x-ray flare intensity. As expected, we found relatively good correlations between the CME speed and the SPE time scales such as rise, duration, and decrease times, which mean that the faster CMEs may drive shocks for a longer period, resulting in the SPEs of longer duration and decrease times. While inspecting the relation between the SPE peak intensity and the direction parameter, we found there are two groups: the first group consists of extremely large six SPEs (proton peak intensity is over 104 pfu); the correlation coefficients is 0.65, which is the highest correlation in this correlation study and there is no one
ARTICLE IN PRESS 360
J. Hwang et al. / Acta Astronautica 67 (2010) 353–361
2001/09/24
2000/07/14
2001/11/22
2001/11/04
2003/10/28
2000/11/08
Fig. 8. Streamer–CME interactions for the large six events.
to find such a high correlation in many previous statistical studies. Such a high correlation coefficient is in contrast with cc=0.01 for the remaining 57 dataset excluding large six events. This clear separation might be caused by different acceleration mechanisms or different acceleration efficiency, so we do investigate the characteristics and solar environments of the large six events in detail. By investigating characteristics of the first group, we found that all the SPEs are associated with very fast halo CME ( 41400km/s). We also found that these large six events were (except for one event) located at disk center within a latitude strip between N22 and S16, and a longitude band between W34 and E23. According to Cane et al. [7], a shape of the SPE time profile is a function of heliolongitude within the framework of a recently derived model for the large-scale structure of IP shocks. In this study, we found the SPE intensities themselves are affected by not only the longitude of the solar event but also its time profile. Therefore, the large six SPEs located at disk center are well-connected to the shock and to the earth, and show the good correlation to the CME earthward direction parameter. In other words, the CME
earthward direction parameter can be an important input parameter of an empirical prediction model for the SPEs arriving at 1 AU, especially if the source longitude is located at disk center and the associated CME is very fast. In this study, we found a strong correlation between the SPE peak intensity and the CME earthward direction parameter for the large six SPEs, while the correlations between all other parameters were found to be marginal or poor. The reasons why the SPE intensity is proportional to the CME earthward direction parameter could be understood by considering the following facts: (1) if we apply the projection effects depending on direction parameter to the six CMEs, a CME with a large direction parameter will gain faster speeds and become an extremely fast CMEs. (2) We may suppose that, there is an IMF connection from a CME front to the earth for the CME with the large direction parameter and that from a CME flank to the earth for the CME with the small direction parameter. The shock ahead of the CME front with the large direction parameter should be stronger than that from the CME flank so that the shock could accelerate particles more efficiently than that from the
ARTICLE IN PRESS J. Hwang et al. / Acta Astronautica 67 (2010) 353–361
CME flank with the small direction parameter. In addition, the particles from the CME front with the large directional parameter could propagate directly along the IMF line connecting the CME front and the earth. (3) Extremely fast CMEs with large direction parameters near the Sun will sustain their speeds and become strong IP shocks during the Sun to the earth. The strong IP shock can accelerate particles continuously and produce strong SPEs. We found that there is the good correlation between the SPE peak intensity and the CME earthward direction parameter only for the large six events while there is no correlation for the low-intensity SPEs. One may have a question what makes a clear separation between those two groups. To answer this question we examined the large six SPEs in detail and noted that there are preceding CMEs originated from the same solar source region, and there exists a nearby helmet streamer. We speculate that the preceding CME and pre-existing helmet streamer might provide seed particles by the following CME-driven shocks. Our results support the requirement for the highflux SPE, which was proposed by Gopalswamy et al. [22], i.e., the primary fast CME is originated from the same active region as that of the preceding CME and interacts with the preceding slow CME. The presence of the preexisting streamer might give the similar effect with the preceding CME. For SPEs in which the primary CME interacted with the streamer or with the preceding CME very close in time, proton intensities were highly correlated with the CME direction parameter. In this respect, a particle acceleration mechanism of SPEs associated with the CME–CME interaction and the CME–streamer interaction needs to be investigated in the near future.
Acknowledgement This work has been supported by the ‘Development of Korean Space Weather Center’ of KASI and the KASI basic research funds. Data of space satellite of NGDC, CDAW and GOES have been utilized. Yong-Jae Moon has been supported by the WCU grant (No. R31-10016) funded by the Korean Ministry of Education, Science and Technology
361
and by the Korea Research Foundation Grant funded by the Korean Government (MOEHRD, Basic Research Promotion Fund) (KRF-2008-314-C00158, 20090071744). References [1] D.V. Reames, Space Sci. Rev 90 (1999) 413. [2] S.W. Kahler, J. Geophys. Res. 106 (2001) 20, 947. [3] N. Gopalswamy, S. Yashiro, M.L. Kaiser, R.A. Howard, J.-L. Bougeret, Astrophys. J. 548 (2001) L91. [4] N. Gopalswamy, S. Yashiro, M.L. Kaiser, R.A. Howard, J.-L. Bougeret, Geophys. Res. Lett. 29 (8) (2002) 1265, doi:10.1029/2001GL013606. [5] N. Gopalswamy, S. Yashiro, A. Lara, M.L. Kaiser, B.J. Thompson, P.T. Gallagher, R.A. Howard, Geophys. Res. Lett. 30 (12) (2003) 8015, doi:10.1029/2002GL016435. [6] M.A.I. van Hollebeke, L.S. Ma Sung, F.B. McDonald, Sol. Phys. 41 (1975) 189. [7] H.V. Cane, D.V. Reames, T.T. von Rosenvinge, J. Geophys. Res. 93 (1988) 9555. [8] C.C. Balch, Radiat. Meas. 30 (1999) 231. [9] D.V. Reames, L.M. Barbier, C.K. Ng, Astrophys. J. 466 (1996) 473. [10] D.V. Reames, S.W. Kahler, C.K. Ng, Astrophys. J. 491 (1997) 414. [11] E.I. Daibog, Yu.I. Logachev, S. Kahler, K. Kecskemety, Adv. Space Res. 32 (12) (2003) 2655. [12] V.G. Stolpovskii, E.I. Daibog, S.W. Kahler, G. Erdos, Adv. Space Res. 21 (1998) 543. [13] N.R. Sheeley Jr., J.H. Walters, Y.-M. Wang, R.A. Howard, J. Geophys. Res. 104 (1999) 24739. [14] Y.-J. Moon, G.S. Choe, H. Wang, Y.D. Park, N. Gopalswamy, G. Yang, S. Yashiro, Astrophys. J. 581 (2002) 694. [15] S.W. Kahler, Astrophys. J 628 (2005) 1014. [16] /http://sec.noaa.govS. [17] /http://cdaw.gsfc.nasa.gov/CME_list/UNIVERSAL/2003_07/ univ2003_07.htmlS. [18] /http://sxi.ngdc.noaa.govS. [19] Y.-J. Moon, K.-S. Cho, M. Dryer, Y.-H. Kim, Su-chan Bong, Jongchul Chae, Y.D. Park, Astrophys. J. 624 (2005) 414. [20] G.M. Mason, J.E. Mazur, J.R. Dwyer, 3HE enhancements in large solar energetic particle events, Astrophys. J. 525 (1999) L133. [21] N. Gopalswamy, S. Yashiro, G. Michaek, M.L. Kaiser, R.A. Howard, D.V. Reames, R. Leske, T. von Rosenvinge, Astrophys. J. 572 (2002) L103–L107. [22] N. Gopalswamy, S. Yashiro, S. Krucker, G. Stenborg, R.A. Howard, J. Geophys. Res. 109 (2004) A12105, doi:10.1029/2004JA010602. [23] N. Gopalswamy, S. Yashiro, G. Michalek, M. L. Kaiser, R. A. Howard, R. Leske, T. von Rosenvinge, D. V. Reames, in: M. Velli, R. Bruno, F. Malara (Eds.), AIP Conf. Proc., 2003, 679, 608. [24] K.-S. Cho, S.-C. Bong, Y.-H. Kim, Y.-J. Moon, M. Dryer, A. Shanmugaraju, J. Lee, Y.D. Park, A&A 491 (2008) 873, doi:10.1051/0004-6361:20079013. [25] R.-S. Kim, K.-S. Cho, K.-H. Kim, Y.-D. Park, Y.-J. Moon, Y. Yi, J. Lee, H. Wang, H. Song, M. Dryer, Astrophys. J. 677 (2008) 1378–1384.