Transient Shocks and Associated Energetic Particle Distributions

Transient Shocks and Associated Energetic Particle Distributions Observed by ACE during Cycle 23 G. C. H o ' , D. Lario', R. B. Decker', C.W. Smlth^ and Q. H u ' "The Johns Hopkins University Applied Physics Laboratory, Laurel, MD 20723, USA Institute for Earth, Oceans, and Space, University of New Hampshire, Durham, NH 03824, USA 'Institute of Geophysics and Planetary Physics, University of California, Riverside, CA 92521, USA Abstract. Characteristics of both interplanetary (IP) shocks and associated energetic particle distributions (e.g., time-intensity profiles, peak intensities, energy spectra indices, etc.) observed near 1 AU during solar active periods are analyzed. From February 1998 through October 2003, the ACE spacecraft (at 1 AU) detected 298 IP shocks (i.e., occurrence rate ~50 IP shocks/year). A subset of 191 IP shocks was identified as transient, fast-forward shocks. We summarize the findings of the statistical survey. Of the 191 sample shocks, 123 (64%) were associated with intensity increases in 47-68 keV ions, 62 (32%) with increases in 1.9-4.8 MeV ions, and 39 (20%) with increases in 38-53 keV electrons, all measured by the EPAM instrument onboard ACE. It is noteworthy that even in the relatively low-energy ion range 47-68 keV, 68 (36%) of the 191 shocks did not produce an observable intensity increase. The 191 shocks were classified according to the temporal evolution of their time-intensity profiles and of their energy spectra indices. Ion spectra indices in the immediate post-shock region of the shocks were compared to spectral indices expected from steady-state diffusive shock-acceleration (DSA) theory. The majority of the measured spectral indices do not agree with the DSA-predicted indices, which depend only upon shock compression ratio. Ion spectra measured at the shock are often softer than the ion spectra measured well upstream of the shock. This suggests that shock interactions of >47 keV ions are weak at 1 AU; consequently, the strong interaction (i.e., multiple shock crossings) description may not apply to most shock events at 1 AU. Keywords: Particle Acceleration; Shock waves. PACS: 90.50.Pw

INTRODUCTION Case studies of interplanetary (IP) shocks and shock-associated particle distributions observed in-situ provide one way to test models of charged particle injection and acceleration at colhsionless shocks. Since models generally consider idealized conditions and ensemble-averaged quantities, their predictions are neither validated nor refuted convincingly by detailed comparisons with individual shock "event." A more fruitful approach, which we adopt in this paper, is to compare model predictions with statistical surveys compiled from observations of many shocks and associated particle distributions [1-4]. This approach requires assemblage of a sufficient number of shock events to make such statistical comparisons meaningful (the term "event" denotes availability of both shock parameters and associated particle CP1039, Particle Acceleration and Transport in the Heliosphere and Beyond— 7^^ Annual Astrophysics Conference edited by G. Li, Q. Hu, O. Verkhoglyadova, G. P. ZanJi, R. P. Lin, and J. Lulimann O 2008 American Institute of Pliysics 978-0-7354-0566-0/08/$23.00

184

Downloaded 09 Sep 2008 to 128.244.244.60. Redistribution subject to AIP license or copyright; see http://proceedings.aip.org/proceedings/cpcr.jsp

data). Fortunately, the relatively high occurrence rate of IP shocks observed near 1 AU during solar active periods (~50 shocks/year during cycle 23) makes this approach possible. We describe the assemblage and analysis of such a base of statistical samples using data from NASA's Advanced Composition Explorer (ACE) spacecraft. Intensity increases of energetic ions associated with the passage of IP shocks are called energetic storm particle (ESP) events [5]. During a typical ESP event, lowenergy ion intensities may increase by orders of magnitude. We shall use the term "classical ESP event" for those events that display gradual, energy-dependent increases both in pre-shock ion intensities and in energy spectra as well as relatively small upstream anisotropics. We use terms such as "shock spikes" or "step-like postshock increases" to describe other types of particle intensity enhancements associated with the passage of shocks. Particle acceleration at shocks, particularly at ESP events, has been studied since first observed in the 1960s [e.g., 1, 2, 6]. Early models tended to invoke transport conditions that were either in the diffusive [7] or scatter-free limits [6]. Jokipii [8] showed that in the diffusive limit, when particle pitch angle distributions are nearly isotropic, diffusive shock-acceleration (DSA) at oblique shocks includes both firstorder Fermi and shock-drift acceleration. ACE is located at the LI Lagrangian point -230 RE upstream of the Earth, ensuring its continued exposure to IP shocks. Based on an earlier ACE data set, Lario et al. [3] reported statistical properties of intensity-time profiles of both energetic ions and electrons at IP shocks observed during September 1997 through December 2001. The present paper uses ACE data acquired during February 1998 through October 2003.

OBSERVATIONS The Electron Proton Alpha Monitor (EPAM) instrument [9] on ACE provides energetic particle intensities associated with IP shocks. EPAM measures both ions (47 keV-4.8 MeV) and electrons (38-315 keV) in several energy channels. Angular information over nearly a full unit-sphere is obtained on the spinning spacecraft by multiple detectors that have different look directions relative to the spacecraft spin axis but cover comparable energy ranges in nearly the same set of energy channels [9]. ACE observed 298 IP shocks from February 1998 through October 2003 (a list of shocks can be found at http://www-ssg.sr.unh.edu/mag/ace/ACEhsts/obs_list.html). Of these 298 IP shocks, we selected 191 fast forward shocks with clear evidence of being driven by, or related to, the passage of Interplanetary CMEs (ICMEs). Thus we have excluded reverse shocks, slow shocks, and shocks associated with other structures such as magnetic holes or stream-stream interactions. For each of the 191 IP shocks in the survey, we analyzed the merged Level 2 (64 s) ACE/SWEPAM and ACE/MAG data to identify intervals upstream and downstream of the shocks when the plasma (density, temperature, vector velocity) and magnetic field data exhibited asymptotic states, which are subsequently verified by the fitted Rankine-Hugoniot (R-H) jump conditions across a shock. We then employed the nonlinear least-squares technique of Szabo [10] to simultaneously solve the complete set of R-H relations for the pairs of upstream and downstream data points and obtain unique values for the various shock parameters.

185

Downloaded 09 Sep 2008 to 128.244.244.60. Redistribution subject to AIP license or copyright; see http://proceedings.aip.org/proceedings/cpcr.jsp

The technique for solving the R-H relations uses a least-squares fitting procedure to determine asymptotic states upstream and downstream of the shock. We have also estimated the upstream to downstream density compression ratio, H, simply by taking the ratio of the averaged plasma density observed four minutes before and after the shock passage (excluding those data points included in the foot and overshoot regions often observed on either side of the shock). In general, the compression ratios r„ (the R-H fit value) and H compare well, with a correlation coefficient above 0.8 [11]. Time-Intensity and Energy Spectra Classification As in the Lario et al. [3] study, we classified energetic particle responses to the passage of IP shocks based on the evolution of both the time-intensity profiles and the energy spectra. Using 1-minute spin-averaged EPAM data, we defined six categories for both the ion and electron time-intensity profiles (see [3] for examples). An additional two years of data and a slightly different time interval do not change the basic findings of Lario et al. [3]. We find: (1) Over 64% (123 events) of IP transient shocks produced intensity enhancements of 47-68 keV ions. (2) Less than 21% (39 events) of IP transient shocks produced intensity increases of electrons > 38 keV. (3) Step-like post-shock increases occur more often in the electron than in the ion intensities. (4) The shock parameters do not determine unequivocally the type of observed time-intensity profile. van Nes et al. [1] found that during the one-hour interval around the shock passage, ion energy spectra in the range 35-1600 keV can be fitted with two power-laws, with a break point near -250 keV. We use the same strategy to analyze ion energy spectra near shock passages at ACE. EPAM's LEMS120 detector covers the ion energy range 47 keV-4.8 MeV in eight differential channels P r - P 8 ' [9]. We fit two power laws to the energy spectra, one covering the lower energies (yi4: 47-320 keV) and the other the higher energies (yss: 320 keV-4.8 MeV). Since the effects of the shocks are more pronounced in the lower energy ion population, we opted to base our spectral classification on yi4 rather than 758 Figure 1 shows the classified by the evolution of 714 for six FIGURE , 1., Six . examples of shock events as,.,, , examples

^ , , o f shock

spectral evolution around the shock. The six diiierent categories events are described in the text.

186

Downloaded 09 Sep 2008 to 128.244.244.60. Redistribution subject to AIP license or copyright; see http://proceedings.aip.org/proceedings/cpcr.jsp

that correspond to each of the six classes in our classification scheme. Based on the evolution of the spectral index yi4 during the one-hour interval centered on the time of the shock passage, we classified the events based on the work of van Nes et al. [1]. We use the four classes (A-D) defined by van Nes et al. [1], and add three more (A*, B*, and E). As in van Nes et al. [1]: class (A) is for events with a gradually increasing yi4, peaking at the shock or sometimes still increasing in the post-shock region; class (B) is for events with several short, strong increases of yi4 that are uncorrelated with the shock passage time but not due to statistical fluctuations; class (C) is for events with short duration (< 10 min) increases of yi4 coincident with the shock; class (D) is for events with a nearly constant yi4 across the shock. We add: class (A*) for events in which yi4 gradually increases just behind the shock; class (B*) for events with several short increases of yi4, with one such increase coincident with the shock; class (E) for events with yi4 constant across the shock because particle intensities remain near instrumental background levels. All but one event in classes (D) and (E) correspond to events with no particle enhancement. Most of the classical ESP events, as defined by the time-intensity profile, have spectral index evolution corresponding to class (A). As expected, all the step-like events are class A*. Note that whereas the time-intensity classification of the events is based on the evolution during a ~10-hour interval of the ion particle intensity in one energy channel (47-58 keV ions) [3], the spectral class is based on the evolution of four energy channels (P' l-P'4) during a one-hour interval about the shock passage. Spectral Comparison 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 111 11 1 11 11 1 111 1 1 1 /

Ho et al. [12] showed the spectral indices for 28 IP shocks that have intensity increases of energetic electrons >38 keV. They found that there are only small changes in the spectral indices for both ions and electrons observed at the shock. Figure 2 shows the spectral index yi4 measured before the onset of the shock event (defined as ambient in Fig. 2) versus yi4 measured at the time of the peak intensity (usually at or shortly after the shock passage). The symbols denote the different spectral classes, and the diagonal dashed line indicates equivalence of the ambient and near-shock spectral index. Normally all class (D) and (E) events he on this line, except when the ambient period is outside the one-hour period used to classify the evolution of yi4. Nearly all

•

•

°

'^ /

m

•

:

°

0

^

o . • *

•

/ ; •

• / . ' 0 /

'. •

J/ •

.•

-

1

^* /

:

• * °* I - V • '

•

•

+ D :

; ; 0

• A : 0 A*: • B :

^

/

0

/

B*:

• c :

r-f '^' 1

1

0

1

1

1 ' '

~

2 3 4 (7i4)ambient FIGURE 2. Two-minute average ion spectral indices before and at shock. Nearly all indices fall above diagonal line, indicating spectral softening at shock passage. Spectral-evolution classification shows no obvious trend. All class D events are along dashed line.

187

Downloaded 09 Sep 2008 to 128.244.244.60. Redistribution subject to AIP license or copyright; see http://proceedings.aip.org/proceedings/cpcr.jsp

events have spectra that are softer at the shock compared to the ambient. However, there is no obvious correlation (except those of class D and E) between the increase in the spectral index y and the classification of the event. Figure 3 shows the spectral index yi4 at the time of the 4 intensity peak versus density compression ratio H for (1) all shocks for which H could be computed, and (2) all particle events of spectral class other than D or E (a total of 107 events). The sohd curve shows y^ versus H as predicted by steady-state DSA theory: y={H+2)/{2H-2), where y is the spectral index. Of the 107 events in Fig. 3, 55 lie outside the area within the dashed curves (DSA theory (solid curve) ± 25%) and 43 outside the area within the longdashed curves (DSA theory (solid curve) ± 40%). 0 .1 1 1 1. Thus, as in Ho et al. [12], 3 5 roughly 50% of the measured 4 H events do not fall within the range of prediction. This is FIGURE 3. Ion spectral index yu of shock events with observable enhancements versus shock compression ratio contrary to what was found using H. Solid curve is DSA prediction y={H+2)l{2H-2), and red ISEE-3 data, where 75% of the (blue) dashed curve indicate a ±25% (±40%) uncertainty of ion events were consistent with the value. the predicted y [1]. van Nes et al. [1] found that the 75% of the events with 35-238 keV ion intensity enhancements and with computed H were within the DSA ± 25% range. Possible reasons for the discrepancy between our results and those of van Nes et al. [1] include the lower energy ion range used by [1], different instrumental sensitivities, and dissimilar event selection criteria. Whereas Figure 12 of [1] contains only 30 events with ion intensity enhancement, our more comprehensive sample of events contains 107 events of diverse characteristics. In addition, the discrepancy between the measured 714 and the y expected from the DSA theory is anticipated if we consider that H is a local (singlepoint) measurement of the shock compression, whereas the intensities of mobile energetic particles, which interact with distant and diverse regions of the shock where H (and other shock properties) varies, carry information of a non-local nature [13].

SUMMARY From 1 February 1998 through 28 October 2003, the ACE spacecraft (1 AU) detected 298 IP shocks (i.e., occurrence rate ~50 IP shocks/year). A subset of 191 IP

188

Downloaded 09 Sep 2008 to 128.244.244.60. Redistribution subject to AIP license or copyright; see http://proceedings.aip.org/proceedings/cpcr.jsp

shocks were identified as transient, fast-forward shocks. Of the 191 sample shocks, 123 (64%) were associated with intensity increases in 47-68 keV ions, 60 (31%) with increases in 1.9-4.8 MeV ion intensities, and 39 (20%) with increases in 38-53 keV electron intensities. It is noteworthy that even in the relatively low-energy ion range 47-68 keV, 68 (36%) of the 191 shocks did not produce an observable intensity increase. We found no dependence on spectral classification of the correlation between ambient upstream spectral indexes and those measured at the shock. Most noteworthy is that almost all of the events have spectra that soften at the shock. Statistical analyses shown herein were motivated, in part, by similar analyses performed for distributions of shock events observed during cycle 21 and reported by [1] and [2]. For example, van Nes et al. [1] reported about 75% of their shock events with 35-238 keV ion intensity enhancements had spectra indices consistent with DSA theory (within a ±25% error margin). In the present paper we find a lower percentage (Figure 3), specifically, only about 50% of the 191 shock-associated 47-321 keV ion spectra indices are consistent with DSA theory.

ACKNOWLEDGMENTS The work at JHU/APL was supported under NASA grants NAG5-13487, NAG510836, and NNG04GA84G.

REFERENCES 1. van Nes, P., R. Reinhard, T.R. Sanderson, and K.-P Wenzel, The energy spectrum of 35- to 1600-keV protons associated witii interplanetary shocks, J. Geophys. Res., 89, 2122-2132, 1984 2. Tsurutani, B.T., and R.P. Lin, Acceleration of >47 keV ions and >2 keV electrons by interplanetary shocks at 1 AU, J. Geophys. Res., 90, 1-11, 1985 3. Lario, D., et al., ACE observations at energetic particles associated with transient interplanetary shocks. Solar Wind Ten: Proceedings of the Tentii International Solar Wind Conference, CP619, edited M. Velli, R. Bruno, and F. Malara, AIP, 640-643, 2003 Kallenrode, M.-B., Particle acceleration at interplanetary shocks - Observations at a few tens of keV vs some tens of MeV, Adv. Space Res., 15, 8-9, 375-384, 1995 5. Bryant, D.A., T.L. Cline, U.D. Desai et al.. Explorer 12 observations of solar cosmic rays and energetic storm particles after tiie solar flare of September 28, 1961, J. Geophys. Res., 67, 4,983, 1962 Decker, R.B., The modulation of low-energy proton distributions by propagating interplanetary shock waves: A numerical simulation, J. Geophys. Res., 86, 4537, 1981 Lee, M . A , Coupled hydromagnetic wave excitation and ion acceleration at interplanetary traveling shocks, J. Geophys. Res., 88, 6109-6119, 1983 Jokipii, J.R., Particle Drift, Diffusion, and Acceleration at Shocks, Astrophys. J., 255, 716-720, 1982 Gold, R.E., S.M. Krimigis, S.E. Hawkins, D.K. Haggerty et al.. Electron proton and alpha monitor on tiie Advanced Composition Explorer Spacecraft, Space Sci. Rev., 86, 541, 1998 10. Szabo, A., An Improved Solution to tiie 'Rankine-Hugoniot' Problem, J. Geophys. Res., 99, 14737-14746, 1994 11. Lario, D., Q. Hu, G.C. Ho et al.. Statistical Properties of Fast Forward Transient Interplanetary Shocks and Associated Energetic Particle Events: ACE Observations, Proc. Solar Winf 11-SOHO 16 "Connecting Sun and Heliosphere", ESA SP-592, 81-86, 2005. 12. Ho, G.C, D. Lario, R.B. Decker et al.. Energetic Electrons Associated witii Transient Interplanetary Shocks: Evidence for Weak Interaction, 28th ICRC, 3689-3692, 2003 13. Giacalone, J., and M. Neugebauer, The Energy Speclrum of Energetic Particles Downstream of Turbulent Collisionless Shocks, ApJ., 673, 629-636, 2008.

189

Downloaded 09 Sep 2008 to 128.244.244.60. Redistribution subject to AIP license or copyright; see http://proceedings.aip.org/proceedings/cpcr.jsp