JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 106, NO. A10, PAGES 20,979-20,983, OCTOBER 1, 2001
Forecasting the arrival of shock-accelerated solar energetic
particles at Earth
C. M. S. Cohen,R. A. Mewaldt, A. C. Cummings,R. A. Leske, and E. C. Stone
Califomia Instituteof Technology,Pasadena,California
P. L. Slocum and M. E. Wiedenbeck
Jet PropulsionLaboratory,Califomia Instituteof Technology.Pasadena,California
E. R. Christianand T. T. von Rosenvinge
NASA GoddardSpaceFlight Center,Greenbelt,Maryland
Abstract. Energeticparticlesacceleratedat interplanetaryshockscanresultin an increased
radiation dose for astronauts as well as an increased risk to satellite hardware.
These shocks
are known to occasionallyaccelerateprotonsto energiesgreaterthan 100 MeV and can
resultin very high fluxes. Advancedwarning of the arrival of stronginterplanetaryshocks
can enablestepsto be takento minimize the potentialrisks. In an effort to monitorand
assessthe radiationrisk to astronauts,two real-time countrate monitorswere implemented
in the Solar IsotopeSpectrometer(SIS) on the ACE spacecraft.Theseratesmeasure
protonswith energies>10 and >30 MeV. SinceACE is locatedat the L1 Lagrangianpoint,
theseratesprovideinformationand warning up to 1 hour beforean interplanetaryshock
reachesEarth. Using ACE and GOES, datawe have examinedexamplesof shocksand
associatedenergeticparticlesthat have been detectedat ACE and were later observednear
Earth in an effort to developmore fully the forecastingcapabilityof ACE. Becauseof the
limited amountof solaractivity during 1997-1999, this currentwork is primarily a proof of
concept.
1. Introduction
Recently, the Committee on Solar and Space Physics
(CSSP) and the Committee on Solar-Terrestrial Relations
(CSTR) examined the risk of increased radiation hazards to
astronautsworking on the International Space Station ([SS).
The results of the study were published in the report
"Radiation and the International
Space Station:
Recommendations
to Reduce Risk"
and indicated
that there is
a significantprobabilitythat severalconstructionflights of the
ISS could be impacted by penetrating particle radiation
events. Of primary concernto the astronautsare events with
significant fluxes of >10 MeV/nucleon particles. During
these events, astronauts could experience a significantly
increasedradiation dose (especially if they are perforlning
extravehicularactivity), possiblyaffecting their future health
and flight schedules.Additionally, increasedexposurecould
requiremore frequentrotationof the crews,therebyimpacting
spacestationschedules
andcosts.
The radiation hazard can be especially high when a
suppression
of the geomagneticcutoff is coincidentwith high
fluxes of energeticparticles [Leske et al., this issue]. The
geomagneticcutoff is suppressed
during geomagneticstorms
when a favorable orientation of the interplanetarymagnetic
field is combinedwith increasedsolar wind pressureon the
Copyright2001 by theAmericanGeophysical
Union.
magnetosphere
[Sheaet al., 1999]. The sameshockswhich
compressthe magnetosphere
in these geomagneticstorms
may acceleratesubstantial
numbersof protons(and heavier
ions) to energies greater than 100 MeV. Such local
acceleration(as well as turbulencenear the shock,which can
confine the particlesto the shock region) can result in
intensities at the shock which temporarily dominate an
ongoingsolarenergetic
particle(SEP) event,aswasthe case
duringthe October19, 1989,SEPevent. Duringthe decline
of this event,the passageof the associatedshockresultedin a
greater than fivefold increase in >100 MeV protons as
measuredby the GOES 7 spacecraft(Figure 1). Since the
time profile of particle intensities in a typical SEP event
decaysrelatively smoothly(especiallyat E > 100 MeV), such
an increase during the decay phase is unexpected and
thereforewould not be predictedwithout information about
the arriving shock. Forehandknowledgeof sucha situation
would allow astronauts
to move to moreprotectedareasof the
ISS or to reschedule
plannedextravehicularactivities.
A largeeffort was made in the 1980sto understandshocks
and their accelerationof particles to MeV energies. It was
found that shockscan be divided into two broad categories,
parallel (where the angle betweenthe shocknormal and the
magneticfield, O•, is 00-45ø) and perpendicular(where Ot•is
45ø-90ø). At parallelshocks,ions gain energythroughFermi
acceleration,a processwhere the ions are reflectedby waves
Papernumber2000JA000216.
upstreamanddownstream
of the shock.Sincethewavefields
are converging,the particlesare being reflectedby moving
0148-0227/01/2000JA000216509.00
barriers and thus gain energy. The more times an ion is
20,979
20,980
COHEN
ET AL.:
BRIEF REPORT
2. Instrumentation
and Data
Selection
The SolarIsotopeSpectrometer
(SIS) on ACE measures
energetic
ionsfrom-10 to 100MeV/nucleon
[Stoneet al.,
1998b]. A stackof large-area
silicondetectors
is usedto
o•
determinethe kinetic energy,nuclearcharge,and massof
incoming
particles.The largegeometry
factorof SIS allows
02
measurements
of low fluxes of heavy ions to be made with
goodtemporalresolution.Thesemeasurements
require
trajectoryinformation
whichis obtainedfromthe top two
detectors,which are positionsensitive. While energetic
protonstypicallydo not depositenoughenergyin these
0ø
ß
detectorsto exceedthe set thresholds(which are optimized
for higher-Zparticles),the energythresholdsof single
detectorsdeeper in the stack are low enoughthat the
Doy of 1989
associated
single-detector
ratesaredominated
by protons.
Shortlybeforelaunch,provisionswere madeto include
Figure 1. GOES 7 fluxes of >10, >30, and >100 MeV protons
-• .....
292
•
294
I
296
I
298
300
two such rates from SIS among those telemetered
duringthe eventsof October 1989.
continuously
aspartof theRealTimeSolarWindmonitoring
systemon ACE [Zwicklet al., 1998]. Thesetworatesarethe
countingratesof the singleT4 detectorandthe coincidence
reflected,the more energyit gains. At perpendicularshocks,
shock drift accelerationis the primary processinvolved. In
this casethe particlesdrift accordingto v x B forces in the
betweenthe T6 and T7 detectors(Figure 2). The T4 and
T6oT7 ratesrespondprimarilyto >10 and>30 MeV protons,
respectively,
similarto the I3 and I4 rateson the GOES
spacecratl.
Increases
in theseSISratesdueto a passing
shock
accelerated.The more time the particlespendsnear the shock
shouldbe followedby corresponding
increases
in the GOES
(under the influence of the electric field), the more energy it
direction
of the induced
electric
field at the shock and are
gains. In both casesthe amount of energy a particle gains
dependson Ot• and the speed of the shock. In shock drift
accelerationthe energy gain also dependson the pitch angle
of the ion. Comprehensiveoverviewsof both typesof shock
accelerationhave been presentedby Armstronget al. [ 1985]
and Scholer[ 1985].
Unfortunately, althoughthe general accelerationprocesses
rates when the shock reaches GOES.
In an effort to study this correlationand develop the
capability
for near-real-time
utilizationof theSIS ratesin a
warningsystem,we haveexaminedthe dataobtainedsince
launchin August1997 throughDecember1999 for shockrelated increases in the T4 rate. An indication that the
increases are a result of local shock acceleration is that both
have been established, the details of shock acceleration of
energetic particles and their subsequenttransport involve a
numberof interplanetaryparameters,and it is hard for particle
acceleration events such as that of October 19, 1989, to be
accurately forecast from observations of the coronal mass
ejections(CMEs) typically driving the shocks[Feynmanand
Gabriel 2000]. It is alsodifficult to predictfrom observations
of eruptingCMEs whethera shockwill developthat will still
be capable of acceleratingsignificant numbersof particles
when it arrives at Earth. Kahler et al. [1984] found a
significantcorrelationbetweenthe speedof CMEs observed
with the Solwind coronagraphand the associatedproton
fluxes measured by the IMP 8 and ISEE 3 spacecraft.
However, for a given CME speedthe observedproton fluxes
varied from event to event by as much as 4 orders of
magnitude[seeKahler et al., 1984; Figure 5].
The Advanced CompositionExplorer (ACE) [Stoneet al.,
1998a], at the L 1 Lagrangianpoint, is far enoughupstreamof
the Earth that it can provide advancedwarnings(---1hour) of
incoming shocks. Since few shocksare strong enough to
accelerate particles to high energies, it is not adequateto
merely observethe shockat L1. A measureof the energetic
particleslocally acceleratedby the shockwhen it passesACE
is needed for a reasonable indication
of what will occur-when
the
Real-Time
shock
reaches
Earth.
The
Solar
'%
II ',
o,.
entrance
75 ",
,'
II
,'• M1
75"
"M2
100
T1
T2
100
250
5OO
13
T4
75O
T5
2650•
1 ooo
,
T8
I
Wind
i
I
cm
monitoring system [Zwickl et al., 1998], 'which is a
compilation of ACE measurementstelemetered in near real
time, providesthe necessaryinformationfor monitoringshock
Figure 2. Schematicof the Solar IsotopeSpectrometer(SIS; one
of two telescopes). Two real-time rates are derived from the
single T4 detector and the coincidence of the T6 and T7
events.
detectors.
COHEN ET AL ß BRIEF REPORT
20,981
the T4 and T6.T? rates rise togetherwith negligibleenergy plottedin Figure3. Shocksidentifiedby the MAG and
dispersion.
Energydispersion
is typicalof SEPeventswith SWEPAM teams(C. Smith, private communication,2000)
acceleration
occurringnearthe Sun,sincethe higher-energy are indicatedby vertical lines.
particles
travelat a greatervelocityandthusarrivefirst. In
addition,we have examinedthe solar wind proton speed 3. Observations
profilesfromthe SolarWindElectron,Proton,andAlpha
As can be seenfrom Figure3, the time profilesof the rate
Monitor(SWEPAM)level2 data(seeMcComaset al. [1998]
increases
observedby SIS and GOES can be remarkably
for the instrument
description)
for abruptincreases
typicalof
a shock. An increase in the I3 rate of GOES was then' similar. Duringthe 1998 day 267 event,manyof the small
searchedfor, and if found, taken as an indication that the variationsin the ratesare apparentin bothdatasets. This is
an indicationof how little the propertiesof the shockhad
shockreachedEarthandwas still acceleratingparticles.
Fivetimeperiodswereidentifiedin theSIS data,butsolar changedduring the propagationfrom L I to the Earth's
windplasmaandmagnetic
fielddatawereonlyavailablefor magnetosphere.Although GOES is rarely outsidethe
thebowshockandmagnetopause
havelittle
three of them. The SIS T4, GOES I3, SWEPAM proton magnetosphere,
velocityanddensity,andmagnetic
fieldstrength
profilesfrom effecton the shockparticlesat theseenergies.
The timedelaysbetweentheparticleincreases
seenat ACE
the Magnetometer
(MAG) (seeSmithet al. [1998] for the
and
GOES
were
determined
by
temporally
shifting
the two
instrument
description)
for thethreeselected
timeperiodsare
102[•...............
10 2
I ................
:•.•'•10'•-
) •
10'
102
½t,o' -
•', 10'
,-
lOOI-I'IP'I
II
I
00
'!
50'
50'
40
•o
•o
30
• •20
20
lO
10
ooo'
000
800
800
400
400
200
2OO
30
60'
20
40
E 15
E
20
I
5 pt•
¾•'•"'• •'•"• •
0
ß ß ß • .
267.2
267.4
,
,
•
,
267.6
,
,
267.8
268.0
268.2
Doy of 1998
ß
268.4
ß ß ß
268.6
268.8
0
123.4
123.6
123.8
124.0
124.2
124.4
Doy of 1998
Figure 3. (a-c) Threeselectedtime periodswhereshock-accelerated
>10 MeV protonswere observedfirst by SIS (first
panel) and then later by GOES (secondpanel). The third throughfifth panelspresentthe solarwind magneticfield
magnitude,protonvelocity, and densityas measuredby the MAG and SWEPAM sensorson ACE during the events.
Shocksidentifiedby the MAG and SWEPAM teamsare indicatedby theverticallines. The list of shocksdid not include
1999 data, so none are indicatedin Figure 3c.
124.6
20,982
COHEN ET AL.: BRIEF REPORT
10 2
flux and momentumflux conservation).The time requiredfor
the shock to traverse the distance between L1 and Earth at the
0,--
calculatedradial speedis given in Table 1 for time periods
1998 day 267 and 1999 day 022. It wasnot possibleto obtain
a shock speed for time period 1998 day 124 since the
simplified equationsused here were not applicableto the
>• •'•o'
conditionsat that shock. The given uncertaintiesindicatethe
variation
102'
in the calculated
shock normal vector from the
differentequationsgiven by Abraham-Shraunerand Yun.
o
>•
o'
0 m
Althoughthereare only two examples,andthe uncertainty
is large on the calculatedtime delay for the 1999 day 022
shock, it appearsthat the time delays observedbetween the
^
SIS and GOES rates are consistent with the calculated radial
E
U') v
shockspeeds.This indicatesthat shock-related
particlefluxes
0
observed by SIS are a reasonable indication of what GOES
25'
ß
'='-
20
:•
1,5
will measure30-60 min later and that the time delay can be
determined from the shock speed as calculated from the
measuredplasmaparameters.
17'-"10
.•_
c
4. Summary
,5
o
800
600
400
200
o
22.4
22.6
22.8
2,3.0
2,3.2
2,3.4
2,3.6
Doy of 1999
With the construction of the International Space Station
(ISS), the radiation environment to which astronauts are
exposed is of particular concern. The assembly of the ISS
will require numerousextravehicular activities during which
astronauts are more susceptible to the increased radiation
possiblefi'om high intensitiesof solarenergeticparticles. The
ability to provide advancedwarning of such conditionscan
greatlyhelp to mitigatethis risk.
In examining time periods in which shock-relatedparticle
increases were observed initially by SIS on ACE and
subsequentlyby GOES, we have shown that warnings of 3060 min are possible. Since the vast majority of shocks
passing ACE are not strong enough to accelerate energetic
particlesto >10 MeV and causea significantincreasein their
intensities, it is not sufficient to merely identify the
occurrence of a shock at ACE. Measurementsof energetic
particlesin real time are also required. SIS providestwo realtime count-ratemonitors which respondto protonsat energies
of >10
Figure 3. (continued)
and
>30
MeV.
These
data
are available
via
the
Internet at http'//sec.noaa.gov/ace/SIS_3d.html
and
http://www.srl.caltech.edu/ACE/A SC.
profiles relative to each other until distinguishingfeatures
The work presentedhere is primarily a "proof of concept."
were best matched by eye. These time delays are given in Additional examples are needed to evaluate the usefulness
Table I alongw•th uncertaintieswhich reflectthe variationof and accuracy of this technique. It is also important to
the determinedtime delays fi'om the different featuresof the determine how many shock-accelerated increases are
rate profiles.
observedby SIS but not by GOES' how many increasesare
Using the proton velocity componentsand densitiesfrom measuredby GOES and not identifiedwith the proposed
the SWEPAM data and the magneticfield componentsfrom technique;as well as how many incidentsare forecast
the MAG data, the shock speedand normal were calculated correctly.
using the equationsgiven by Abraham-Shraunerand Yun
[1976]. The normal plane is definedas the planethat contains
the magnetic field vectors(ahead of and behind the shock)
Table 1. ShockTime Delays
'•
Time
Period
SIS-GOES
AT.
of the shock. Equationsinvolving five differentcombinations
and the vector difference
of the flow velocities
on both sides
Year
Day
Time,
rain
of these vectors can be used to define this plane, and a
UT
consistencycheck betweenthe different expressionsprovides
1998
267
2330
43 = 7
an indication of the reliability of the normal vector
1998
124
0209
36- 10
determination. The shock speedis then calculatedusing the
1999
022
2120
66 = 5
reduced Rankine-Hugoniot conservationequations(which
%IS, SolarIsotopeSpectrometer.
combine Maxwell's equationswith the assumptionof mass
ShockDerived
AT,
min
38 ñ 3
69 z=30
COHEN
ET AL.:
BRIEF REPORT
20,983
Acknowledgments. We thank the SWEPAM and MAG teams McComas, D. J., et ai., Solar Wind Electron Proton Alpha Monitor
and the ACE ScienceCenter [or providing level 2 data for shock
(SWEPAM) for the AdvancedCompositionExplorer, SpaceSci.
Rev., 86, 563-612, 1998.
analysis. We thank Adam Szabofor helpful discussions
regarding
the shockcalculations. The GOES data were contributedby the U.S. Scholer, M., Diffusive acceleration, in Collisionless Shocks in the
Heliosphere: Reviewsof Current Research, Geophys.Monagr
Departmentof Commerce,NOAA, SpaceEnvironmentCenter. This
work was supported by NASA at the California Institute of
Ser., vol. 35, editedby B. T. Tsurutaniand R. G. Stone,pp. 287301, AGU, Washington,
D.C., 1985.
Technology (under grant NAG5-6912), the Jet Propulsion
Shea, M. A., D. F. Smart, and G. Siscoe, Statistical distribution of
Laboratory,and the GoddardSpaceFlight Center.
geomagnetic
activitylevelsasa functionof solarprotonintensity
JanetG. LuhmannthanksStephenW. Kahler and Ronald D. Zwickl
at Earth, Proc. Con/.'. CosmicRays 25th. 7. 390-393, 1999.
for their assistancein evaluatingthis paper.
Siscoe, G. L., et al., Radiation and the h•ternational Space Station:
Recommendationsto Reduce Risk, 76 pp., National Academy
Press,Washington,D.C., 2000.
References
Smith, C. W., et al., The ACE Magnetic Fields Experiment, Space
Abraham-Shrauner,B. and S. H. Yun, Interplanetaryshocksseenby
Sci. Rev., 86. 613-632, 1998.
Ames plasmaprobe on Pioneer6 and 7, J. Geophys.Res., 81, Stone,E. C., et al., The AdvancedCompositionExplorer,SpaceSci.
2097-2102, 1976.
Rev., 86. 1-22, 1998a.
Armstrong,T. P, M. E. Pesses,and R. B. Decker, Shock drift
Stone,E. C., et al., The Solar IsotopeSpectrometerfor the Advanced
acceleration,in Collisio•lessShocksin the Heliosphere: Reviews
CompositionExplorer,SpaceSci. Rev., 86. 357-408, 1998b.
of CurrentResearch,GeopltvsMonagrSet'.,vol. 35, editedby B. Zwickl, R. D., et al.. The NOAA real-time solar wind (RTSW)
T. Tsurutaniand R. G. Stone, pp. 271-285, AGU, Washington,
systemusingACE data, SpaceSci. Rev.. 86. 633-648, 1998.
D.C., 1985.
Feynman,J. and S. B. Gabriel.On spaceweatherconsequences
and
E. R. Christian and T. T. yon Rosenvinge,Code 661.0, NASA
predictions,
J. Geophys.
Res..105. 10,534-10,564,2000
GoddardSpaceFlight Center,Greenbelt,MD 20771.
Kahler, S. W., et al., Associationsbetweencoronal massejections
C. M. S. Cohen, A. C. Cummings, R. A. Leske, R. A. Mewaldt,
and solar energeticproton events,J. Geoph?s.Res., 89, 9683- and E. C. Stone, MC 220-47, California Institute of Technology,
9693, 1984.
Leske, R. A., R. A. Mewaldt, E. C. Stone,and T.T. von Rosenvinge,
Pasadena,CA 91125. (cohen((v•srl.caltech.edu)
P. L. Slocum, M. E. Wiedenbeck, M/S 169-327, Jet Propulsion
Laboratory,4800 Oak Grove Dr., Pasadena,CA 91109.
Observationsof geomagneticcutoff variations during solar
energetic particle events and implications for the radiation
June30, 2000; revisedOctober25, 2000;
environmentat the SpaceStation,J. Geophys.Res., this volume, (Received
2000.
accepted
November17, 2000.)