AD--A 2 3 1 453 ~ A

$LrLu~i DOCUMENTATION PAGE O1fo014j i

eohSic Labrtr H

ilasr ifL

2t)EVJ JAN N199 "unlimited

'5 02 2311C)tN RtP G4rL%I r.w S~nyC;,, Uctcl

GoLaR9034 Prtn1et3uigteatTreSlrCce

,,a iSOAL orA..11A1: F rFC rIO aN M FM~iu?%,LruVi'Tj

D.F~I Smart, .A. She

M8sachusets 013-500

3" / 4V C, , S ". "Q i,> 183.'J Gii CT f%1t 9l' (Canti Ck HM NTr.e520(.C1JyJhJ.ljfrl ytik1lftFEA 1.,P (It7)U Cosmic rays emgetcsom ageopee

6, (Magnetospheric 1 curen.ts Osmi rayA cutofft riidt

* ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ k( 3A A.uj, i Co~~ ilf~(I ttitlaym Ont yOu. i~ttf

N~~~~~ur;%,Fj NO~i jW Co~ rlnetII ibufuxo ~r1 a~d~(m 2~ ) % b~ 0 ~Vh~

Solar Proton c Event Duingo ~aasuaait the Pas Tre Solar celirstdrgi.

~~~DF mrM.A. Shea 613-'g2'

1989, Page 403 i 415

C j , AT s - ,; 8 L6A T T M Coninueor e~ese t nee( ry t;Ljidtntiy b blok 1Lfrb1f

GL-TR-90-,0346

Solar Proton Events During the Past ThreeSolar CyclesD. F. Smart and M. A. Shea

Reprinted from

Journal of Spacecraft and RocketsVolume 26, Number 6, November-December 1989, Pages 403-415

AMERICAN INSTITUTE OF AERONAUTICS AND ASTRONAUTICS. INC.370 L'ENFANT PROMENADE, SW *WASHINGTON, DC 20024

VOL. 26, NO. 6. NOV.-DEC. 1989 J. SPACECRAFT 403

Solar Proton Events During the Past Three Solar Cycles

D. F. Smart* and M. A. SheatAir Force Geophysics Laboratory, Hanscom Air Force Base, Bedford, Massachusetts

More than 200 solar proton events with a flux of over 10 particles (cm2-s-sr) - 1 above 10 MeV have been rec-orded at the Earth since 1955; at least 151s of these events had protons with energies >450 MeV. Although themajority of solar proton events occur in the years around sunspot maximum, these events, including those withrelativis'ic protons, also occur during sunspot minimum. Detector technology has evolved during the last threesolar cvces so that a uniform data base does not exist. A count of the number of solar proton events above a spe-cific flux level does not indicate a recognizable trend over the three solar cycles. However, an analysis of the par-ticle fluer 'e in "'major" events indicates that the 19th solar cycle was the most energetic.

Introduction volume to change the state of a circuit element thereby causingT HE fact that the suii can accelerate particles to high a "soft error." Also, heavy cosmic-ray nuclei may cause dis-

energies has been known for approximately 40 years. placements in the crystal structure and permanent damage. AnSolar particle events are referred to by a number of dif- examination of the effects of cosmic radiation on microelec-ferent names such as solar cosmic-ray events, solar proton tronics is given by Adams et al. 2

events, solar electron events, polar cap absorption events, and The energies of the cosmic radiation are normally expressedground-level events (GLE). To describe the solar particle in units of GeV (101 eV). A I-GeV proton can penetrate ap-events that have occurred 01iring the past three solar cycles, it proximately 400 g-cm 2 of material, therefore shieldingis necessary to understand several aspects of cosmic radiation against galactic cosmic radiation in space is generally ineffec-such as the background radiation in space and solar particle tive. The primary cosmic radiation observed at the Earth's or-propagation. Consequently, the first part of this paper will bit consists of approximately 83% protons, 13076 alphas, 1%describe the galactic cosmic radiation and its solar modula- nuclei with atomic number Z>2, and 3% electrons. This com-tion; the solar particle events during the last three solar cycles position extends over an energy range from approximately 10are discussed in later sections. MeV to > 102) eV. A more detailed review oriented toward en-

vironmental specification is given by Smart and Shea.3

Charged Particle Background Radiation Solar Cycle of Galactic Modulation Cosmic RadiationThere is a cosmic-radiation background present everywhere The Earth is located deep within the heliosphere (the do-

in space, the intensity being dependent upon the measurement main controlled by solar emissions of plasma and magneticlocation. Spacecraft outside the Earth's magnetosphere are field), which extends to perhaps several hundred times the sun-constantly subjected to the background galactic cosmic radia- Earth distance. (The mean distance from the sun to the Earthtion upon which the solar cosmic rays are episodically super- is defined as an astronomical unit, or a.u.) Since cosmic-rayposed. Spacecraft operating within the magnetosphere are particles originate outside the heliosphere, they must "work"subjected to this radiation; however, the shielding effects ofthe Earth's magnetic field reduce the total flux impinging tary magnetic ield if they are to arrive in the inner solarupon the satellite by an amount that depends on the character- t em. magneti c e in the a netarr i m i s a i n co nistcs f te sacerat obit Saelltesinthepolr rgios wll system. Turbulence in the interplanetary medium is a functionistics of the spacecraft orbit. Satellites in the polar regions will of solar activity; therefore, the cosmic-ray flux at the Earth isbe subjected to the full intensity of galactic and solar particle a function of the solar cycle. Since the solar activity increasesflux; those in the equatorial regions will be subjected to lesser the turbulence that the cosmic ray must "work" against, theamounts depending on altitude. In addition to the galactic and minimum cosmic-ray flux is observed with the solar activitysolar cosmic-ray flux, spacecraft operating within the magne- maximum. (There is an approximate seven-month lag betweentosphere also encounter the trapped particle radiation. solar activity and cosmic-ray modulation that is interpreted as

(Galaclic Cosmic Radiation the time it takes the solar plasma to propagate out into the dis-tant heliosphere.) Similarly, the cosmic rays perform a mini-

The galactic cosmic radiation (whose origin is still a matter mum amount of "work" against a quiet solar wind, so theof scientific debate) is composed of atomic nuclei that have cosmic-ray flux is maximum during solar minimum condi-been ionized and then accelerated to very high energies. The tions. Models exist that will predict the cosmic-ray flux as adeposition of this energy can affect the material through function of the cosmic-ray modulation parameter4 .5 describedwhich the cosmic ray passes. For small, state-of-the-art, solid- as a heliocentric electric field having a potential equal to thestate electronic devices, the passage of cosmic-ray particles energy that arriving cosmic rays have lost. The cosmic-raythrough !hc device can produce enough charge in the sensitive modulation parameter has not been successfully predicted,

and proxies such as solar sunspot number predictions are used(see review of sunspot number predictions by Withbroe6 ). Anexample of using the sunspot number as a cosmic-ray intensitypredictor is given by Adams et al. 2 and Adams.'Re.ced teb. 8, 1989: presented at the AIAA Aerospace Engineer- The solar cycle modulation of the galactic cosmic radiation

ig ( onterence and Show, Ios Angeles, (A, Feb. 14-16, 199; revi- observed at the Earth's surface by a neutron monitor is shownsiion receced April 27, 1999. this Paper is declared a work of the U S.(w,ernmeni and is nolt subject to copyright protection in the Uniled in Fig. I. This figure illustrates that, within a small uncer-states, tainty, the cosmic-ray intensity has been approximately the

*Research Physicist, Space Physics Division. same for each soar minimum for which we have measure-tRescarch Physicisl. Space Physics l)ivision, Member AIAA ments. The extent of the solar modulation on the cosmic-ray

404 D. F. SMART AND M. A. 5HFA J. SPACECRAFT

- - 0

Fag. I_ Solar cycl modulaion- of cI cII- ry80 13

thtpntaeono h eishr owrIh u eo e

T 92-iE--0-90

n66- (l8

C H- 84

C I io ize ytheslrE Veiso.Teenwycetdin'4" lertio prcese.-o -reet eie-f th -nma8 u

11g 1 olaSolarmodlatonoftosmc res

enere nar 10 the paer nuien This phnomaous obsmedic-; io he oay comotpwihr rextns from att -7 Mude pero

,-ROGENenuclaeo, ouissnow expaneras folos: Inesteafr neutalea

" F . 21- ._. ,\thate eetrates into theneSspere codtoarte sunorbeomesi ;/ / \ iomnieytheslac U emission .nt frmthesie eney reeaos

1o_4 " inapreithe eue eb poolyhounderstod ocainerntary acce-lrion rocesss. o a rcetr e vpoiw to the anmalousosmic ray ooa n et s ee -ry e ibonbe vbeatIau.in

-o r ' ' Solar laen d ist the lnamet ien totph enme non obs; ' T eeda in• protesa ywchophrie re elg afed the sdden cv releseo

IO~- IN erate elcos andll ionse , and, i cn ditirof en ss are rbein-

I jetdtheseA presnto sae . eerha ll c habroadctrisi edcitoantieismusioemcanatfrm thayhve sitfeerlae

oo4 o§ ' in oo o0 sagpeadablerqeathoghnis ocasion, etines-osions doae toga first-orderappoiain.h nest

KIN TI ofRY GVnu)A te radio and f x-ray emissions osale aeted aithu. isin

heliu, andironnucledepend weno the loatino h flare n h visi bte xa c k t

ote i etp s fe iso o mal x etd wt so la

v pflressb c particl esae acceleratio e d ikvn thfor activeficre-

gona i satil debede a a nube opf mcns are beng10e y r 1 . y nstie at paresent, the ecansl s canob ouad cola ssi

e isnrle, t c at o e iplie mehasr mts otht ema heu ralsshock p aes nthrou the Eart as c orh a. o ari n beome s

beeNETICtENERGdYby fiieizo tA t hti e s, r andx-ray emissions ee ecly cated io

oeigp. 2 he C -ry dioew ea energy pectrumaor hydkgrogen, corre snd ngie npica or a spefoont hlat usare-ll

heleium, a iron nuc lcto ot rncln)occursngyet he solfiart flar isrbehn tevisen solavrsk t

oterationm tpse o aremssn rmewall the xpectled wioaflareag -paiebacSroun particles aceertion rma berwe e e s an

os a tolar partie eets .ae gespt the poper einohatn renr-

diffrenialenegy pecrumat au.forhydoge, hliu, gticsoared prtl satre wy ai a t ed owith lrgei esa

enerymanges fomcrCom on1 e TehdonflxasiaoteEt associated with a disappeerng filaent ndh

beng utplied bie o that 2the olate petando n thot we c a aomping fma te ddene.ed ooverlapn The erlow energy prticle-ra diaion background Ter rge s iouns ito ener t ati Ts e s lr fare il

(enrgis o a ewmegeletrovltspernceo)istongy renierate Soar finlansa pro on e vetomget.c emoweve, twoelinfuece byO soaciiyadcnesiyvr yafco f reativelyctood" andios ave afcndii aprimatelyb0.75cor

todung a the sarotations podis opont elt the "big flare syndrome" ad e y haperi whothe chrged-prti b agr d istro ngl ap cted wih pea e mi o r spos em nt Durin the late f er gs ande

ID-4 "U-saped"al seqce sgalthu , n the cpa in o eraio emis-Anoalos Csmi-Ra Coponntsions asocinate woito r pro aeve nts her intet

During~ ~~~~~~~~o the delnrfte2t oa ylaoaiswr 90sadi reeantio ofra hesin aevaiable atIa~ identiie-

nomeon as clle the"anmalos csmicraycompnen"nten bi thae sycindrome foraed byi Kac uhe whocsinc thpopstoreitdfomte"xetd omc oes tyhat h martisolare emisoneray prte sa ante

ray composition; specific elements such as helium and carbon various radio emanations) were usually associated with "signi-were perhaps as much as I0 times "normal" abundances at ficant" solar flares, and it was unclear whether or not any par-

Fig ii CoIIl ilr/y difrnialenrgy specrum- for hyroen corsodn optca emsin peo eo--a sal

Nso% .- IW .)11P1~9 PRO IO( N I \ EN I R II I H L PA*S [ THi RHF:I SOL AR (Y(] [S40

(Icukir 111-01 cto ti nlici\ tidltcotcd - olar particlcs. Nlorc re- penldcut phases: I ) dif-fuiion in ithle solar corona and 2) t!rans-,cnt %%,, 1t k 1,1i ICIItol ccd (Iii conclusloill port into the hehiosphiere along the interplanetary magnetic-

field litnes. Both phases arc illustrated in Fig, 5. I-he coronal'solar Par icle Propavahio propagation distance is indicated in Fig. 5 by the heavN arc on

Jo tindLci stand 111 solar1 part1k!L Llati hase aCCIUllCd ditiriro', the suni.he past h lcc kdLii %,Ic-, one ntist d~c,,rihe solar particle [hle miasmunl particle f lux is assumed to occur at the solar

ipropaoat jot Il c cslki aice rct.h th whbit of file flare site %k ith a gradient ex\tending in the corona from the flareIarti hi a Itc% 110tlitcs It hci, parin:csa~ nter %cr high ecr- site ito other heliolongiitdes. This gradient attenuates the max-

II %tinl hours 1ti 1C 0uQ -e2icr~L\ part1ices. EnAhanced il1111 particle intensit% as theanurditcefo thfle

solar iaii Lii)l p aatstthe larifi ss thiri one or isso 'ite increases. Thiere has been esidence for such a gradientja aId lli'saio 114d LC-ottiauctic disturbanices s Itose since thle early sx stematic anali sis of' spacecraft observations

to t CpCttJs oI) 11C lot erplaict ar% plasmia and field of' solar particles. Alt hough observations suggest that thecltta~ci~%c sliu 1;,c Ipliasta arrises and interacts ss i, the :adient varies betisseen e% entts, a decrease ef a factor of 10 per

Jar i\ tacctopicrc 1IclrcI 1 ll11ti~taICs tfte relatie time C~ rad is reasonable for estimating the solar particle flux atarrls alanA IOTttTl ot soaca:Ieeission at tfte IFarth. heliolonit tides awaN from the flare site. As the particles, dif'-

I tlkc solaricr toatci adiatiott. both thie onset titme fuIse thr1oitgf the solar corona theN are transported into thetAI inisuttilit:l ittcr'it\ o1 (titc solar1 particle flux depend oi fieliospltere along the irittrplanctary magnetic-field line.hc Ilcloloncit Ac of te tlarc s tith respcct to Ile detection lo-

~ II in lpa. I hi' ir ctI1,ionlty reuthecatise particleso ill tttos itos! c5115 alonte itttrplartetar% inagneti-field ARCHIMEDEAN

Arct:ioili tI t llcp1 ttctar\mgei- edtpt is dleter- SUN SPIRAL PAI H

t ut1i h% (ti us sol i l OL0s% riA utifsad the rotation (it the sun------------------------lAT%011,1 inIFtc '0itict11 :dttors la be apprOns mlated by anll SUN -EARTH LINE-\rciuIcdLL1, itt spra ii illustrated ini I ic-. 4. By evintint solarpioioti dat~i k :ant cncrali/e and separate the propagation FLARE

o)I sltt tproton-, Irom tw Hltrc site ito thle arth into two indc Fig. 5 Illustration of the propagation concept.

START OFINCREASE

S'>SOLAR TIME OFFLARE MAXIMUM

RFLATIVE TIME CaI;AMPLITUDE OFOF SOLAR MAXIMUM FLUXPARTIClE 1A?

EMISSIONS.REAIETM

TIME FROMPROPAGATION START TO MAX

DELAY

/ \ Fig. 6 General characteristics of solar proton events observed at I

/ / \ RELATIVE SOLAR PROTON7 IAMPLITUDE VS

IQ T %I SPACECRAFT POSITION

CAY DAYS *E

F . lime scates of solar emissions arrival at I au.

48-9 AUG 1970 1

- N0

...0- .. K.). .A 1 ,

406 1). I. SMART AN) NI. A. SHEA J. SPACECRAFT

I able I Normali/ed elemental abundances of solar energetic particle events

I eV:-':" 1 20 NleV1 - to meV 2 6.7-15 Me\'"

I'ncrgL Range -nerg. Range Energy Range Energy Range

I !1 1.0 L.1 1.0 1.02 lie 2.2 1 2 1.5 E 2 1.5 E-2

I 1 1.0 E 7 4.8 E-8 2.8 E-6A Be 1.5 1-7 6.0 E-9 1.4 E-7

B 1.5 1-7 1.2 1 - 1.4 E-7

, i.(, 4 !.2 E 4 9.6 F-5 1.3 E-4N I N '-5 2.8 1-5 2.7 F-5 3.7 E-5

S( ".2 - 4 2.2 L-4 2.2 E-4 2.8 E-4S. 4.3 1:-7 1.0 L-8 1.4 E-7

W Ne 5I L, 3.5 1-5 3.1 E-5 3.6 E-5

II Na 1.6 1 6 3.5 E-6 2.6 E-6 2.4 E-612 M 4.8 1- 5 3.9 U-5 4.3 E- 5.2 E--5I3 AI A.5 [ 6 3.5 E-6 3.1 E-6 3.3 E-614 Si 3. 8 1- 5 2.8 L5 3.5 E-5 4.2 E-5I P 2.1 I-7 4.3 L-7 1.7 E-7 4.0 E-7

16 S 1.8 L1 5 5.7 F-6 7.8 E-6 6.5 E-6I- ( 1 1,7 [.-7 7.1 E-8IS A r 3.9 1 6 8.7 L-7 7.3 E-7 4.6 E-619 K I. 1 -

7 1.0 E-7

21 (Ca 2.3 t-6 2.6 F-6 3.1 E-6 3.2 E-6

21 ,w 7.8 E-922 i 1.0 -- 7 1.2 E-723 V 1.2 E-824 (r 5.7 F-7 5.0 E-725 Mn 4.2 L-7 1.8 E-726 1-e 4.1 L S 3.3 F-5 3.4 E-52

7 C( 1.0 11-' 4.8 E-7

28 Ni 2.2 L-6 1.2 E-629 ( u 1.4 E-830 Zn 3.8 E-8

Particle events at I a.u. exhibit the following char icteristics: possible particle intensity. An observer whose interplanetarya delay from the solar flare time until the first particles are de- magnetic-field connection is at some other heliocentric loca-tected, a relatisely rapid rise in intensity to a maximum value, tion would observe a flux that has been attenuated by propa-and a sloA, decay to the background level. In general, particle gation through the coronal gradient between the flare positioneents from the eastern hemisphere of the sun have slower and the foot point of the Archimedean spiral path from therates of rise than e'ents from flares west of central meridian, sun to the detection position in space. The onset time of theAlthough the shape of an event may be distorted by features in particle event, the time of maximum intensity, the maximumthe interplanetary medium at the time of the solar particle (peak) particle flux, and the total fluence of the event are alleent, or the particle flux may be considerably modified by functions of the location of the flare with respect to the obser-multiple injections or interplanetary shocks, the general fea- vation location in space and the interplanetary conditions attures of the solar proton time-intensity profile, shown in Fig. the time of the flare.6, are always recognizable. This profile is characteristic of theinner heliosphere; at distances beyond about 5 a.u. all solar Composition of Solar Particle Eventsparticle events are diffusive in character. However, distinct Solar energetic particles are assumed to be acceleratedsolar flare particle increases have been observed by the most above solar active regions fiom the available coronal materialdistant spacecraft, now out more than 40 a.u. daring solar flare events. The acceleration site is apparently

At times, major solar flares populate the entire inner high in the solar corona; studies have concluded that the accel-heliosphere with particles, as illustrated in Fig. 7 where the erated ions pass through less than 30 mg cm 2 of material be-vertical bars indicate the relative flux increase at each space- tween the acceleration site and the observation site in the in-craft. On August 8 and 9, 1970, particle increases on the Pio- terplanetary medium. Observations show variations in the ionneer 8 and 9 space probes together with the small increase on abundances with the hydrogen-to-helium ratios being the mostthe Earth-orbiting IMP 5 satellite were not associated with any variable. The elemental abundance ratios seem to have a slightsolar activity on the visible hemisphere of the sun; however, variation according to the energy of the measurement, andactive region 10882, which produced particle events on August small solar cosmic-ray events have the greatest variability in el-13 and 14, 1970, was on the invisible hemisphere of the sun emental composition. (See Refs. 24 and 25 for recent views onabout three days before east limb passage. 2' A flare source lo- the elemental abundances in solar particle -vents.) When ele-cated approximately 400 behind the east limb is consistent mental abundances observed in large solar energetic particlewith the large increase on August 8 observed by Pioneer 9, the events are plotted vs the first ionization potential of each ele-smaller increase of flux observed on August 9 by Pioneer 8, ment, a depletion of the solar particle abundances is found forand the even smaller increase observed at the Earth. elements with first ionization potential above 10 eV. When ele-

To summarize the "classical" or expected solar particle mental abundances are ordered by the fir- ionization poten-propagation characteristics, an observer who is connected via tial, elements with both high and low first ionization potentialthe interplanetary magnetic-field line to the heliographic loca- are consistent with known coronal abundances, indicating thattion of the flaring region will generally observe the maximum these ions were accelerated out of normal cronal material.

NOV.-DEC. 1989 PROTON EVEN FS W-)RING, IHE PAS[ THRI+ Sol AR CYCLES 407

The same principles insolved for organizing proton (ions vanccs in nuclear physics, more scnsiti,, instruments "', ere de-with Z 1) data also apply to heary ions since the same princi- eloped that could directl. measure the incident particles.pies of corona! and interplanetary propagat;n should apply !hose detecitws cie initially carried b% balloons to ge aboseto all ions independent of mass or atomic charge. Unfortu- as much of the [artI', atmospheric shield as possible; laternately, most of the solar particle data sets currently aailable these detectors sacrc adapted for the initial man-made Earth-are for protons. In order to estimate the probable heavy-ion orbiting su'tellites.fluence from the proton fluence, we have derived tables (see Wkhile cosmic-ray researchers weie deseloping their instru-Table 1) of solar particle event element abundance ratios nor- mentation, high-frequcuCs commnuication engineers, particu-'ealized to hydrogen based on comprehensive spacecraft stu- larly those inolked in the propagation of electromagnetic sig-dies. 2 ,2

nals in the polar regions, noted interference that seemed to beassociated with solar actiits. It is no\, known that chargedHistorical Summary of Solar Particle Event particles interacting with the Earth's iotosphere enhance the

Detection Techniques ionization and change the electromagnetic propagation char-A number of different detection techniques have been used acteristic-, of the medium. In the late 1950's. the development

to detect solar cosmic-ray events, and Fig. 8 illustrates the eso- of the riometer (for radio ionosphere opacity tmeter) prosedlution of detection energy thresholds and detector techniques. to be very sensitise to particle deposition in the ionosphereThe thickness of the lines indicates the relatise number of each directls above tile instrument. Lsen though the riotneter couldtype of detector in use. The differences in shading in the not uniquely distinguish the type of particle, its sensitivits %%asionospheric section indicate changes in detection technique. equivalent to the earlN satellite instrtiments. Most of the solarThe first instances where the sun was unambiguously identi- particle flux and flu,,,'- data -is ailab!c from the 19th solar c',-fied as the source of particles detected at the Earth ssere on cl %%ere dcrised from iometer measirements in the Earth'sFebruary 28 and March 7, 1942. The measurements of the polar regions. Esen no"s the ionosphere call still be used as asolar activity (observed as interference in detection and sur- sery, sen, itise (btt nonlinear) particle detection niedium, since,eillance equipment) "sere shrouded in secrecy by the antago- very loss frequency phase and amplitude changes alongnists of the Second World War." It was not until July 25, transpolar propagation path., haxe the same approximate de-1946, and Nosember 19, 1949, that similar events occurred tection thresholds as particle detectors on spacecraft.and the explanation of solar flare accelerated particles beingdetected on the Earth was given respectable scientific cre- Solar Proton Events: i955-1986dence." ' 2 The initial observations of "solar cosmic rays'relied on measurements of secondary particles generated at the I-Asting Data Hase"top" of the Earth's atmosphere. The original ionization As call be seen from inspection of Eig. 8. there are Earth-chambers and counter telescopes are now classed as muon de- based measurements of solar flare generated particles sincetectors; these detectors respond to high-energy (>4 GeV) pro- 1942. Hosvever, the indirect detection techniques did not stabi-tons interacting at the "top" of the atmosphere. In the 1950's, liZC until approximately 1958, and the spacecraft measure-development of the cosmic-ray neutrn monitor" lowered the mients s"ere not really svstematic until about 1965. Based ottdetection threshold to >450 MeV protons interacting at the contemporary knowledge, it is nossible to interpret the"top" of the atmosphere. A number of these neutron moni- ionospheric-sensed data to form a useful data base extendingtors were deployed for the International Geophysical Year back until about 1955. Inclusion of all of the available ground-(IGY), and many neutron monitors are still operating based and satellite-sensed measurements form a data base ex-although the design has evolved with the development of the tending oser three solar cycles. The data base for solar cycle 19so-called "super" neutron monitor." Concurrently, with ad- is primarily derived from ionospheric data supplemented with

limited spacecraft data in the early 1960's. The data base foisolar cycle 20 is deriked from Earth-orbiting satellite measure-ments of solar particle events plus simultaneous ionosphericmeasirements that allow a cross calibration of the detection

7r TE R5 techiiiques. Finally, the data base for solar cycle 21 is derivedprimarily from very sensitive spacecraft instruments (so sensi-

> '0cKv, tive, in fact, that they "saturate" in large events), with only a%EjTRCN . minor contribution from ionospheric data since many of the

Rpolar ionospheric monitoring stations had closed. There havebeen a number of attempts to assemble available solar proton

- o>- data into catalogs.' "' The National Oceanic and Atmos-pheric Administration/ U.S. Air Force Space EnvironmentServices Center in Boulder, Colorado, maintains a current list

3, CN S . - of solar proton events; see Ref. 41.C ,.- Until a reliable direct-measurement (satellite sensed) solar

-. SATE~.,TEN particle data base s"as available, there were alwsays questions

as to the accuracy of the early data bases and the magnitude ofcontamination by local magnutospheric effects. It is not possi-

,on: ble to assemble a completely homogeneous list of solar protonevents detected over the last three solar cycles, primarily be-

- ONOSP-ER.C cause of the different measurement techniques used. The most, -_ -j SPHERIC homogene-, c!.i1, set aailable that has beeh, deiled fiom a

o - standard observational technique are the "ground-levelX DETECTION \ ,- events" (sometimes called relativistic solar proton events) de-z T EC HNiIJES tted by neutron monitors, since the sensitivity of this instru-ment has been essentially unchanged since its inception in

- ,-, - 9 j J 1953. During the 19th solar cycle, some small ground-level1930 1940 1950 1960 1970 1980 events may not have been identified because of a sparsity of

YEAR detectors. Since the early i')60',, a denser net of instrumentsFig. 8 Conceptual history of the detection thresholds of solar has been installed, particularly in the Earth's polar regions,cosmic-ray events, and events with increases of only a few percent have been

408 D. F. SMART AND M. A. SHEA J. SPACECRAFT

RELATIVISTIC SOLAR PkOTU1ON EVENTS

FRLcJENCY SOURCE ROiTEN

SUN

0 E +4o0W

6-i

2 60 0

T h55 60 65 70 75 8C) 65

YE4R

Fig, 9 High-energy solar proton events for three solar cycles.

readily identified since 1966. Figure 9 shows the distribution equivalent to a l-dB polar cap absorption event. Our primary.of these relativistic solar proton events over the past three data sources were Refs. 39 and 40, and data assembled bysolar cycles; the top part of the figure shows the smoothed Shea and Smart. 42 .4 1 Measurements from the synchronous or-sunspot number; the boLtom part of this figure shows the biting GOES spacecraft 41 %%ere used for the years 1976-1986.number of high-energy solar particle events (GLE events) each Table 2 lists the number of discrete solar proton events withyear; and the right part of the figure shows the location of the a flux > 10 particles (cm 2-s-sr) Ias a function of month forsource solar flare on the sun. An examination of Fig. 9 (at the past three solar cycles (1955-1986). The yearly sums areleast to our prejudiced eye) does not indicate any outstanding shown in the top portion of Fig. 10, and the yearly averagetrends other than the general association with solar activity, sunspot numbers are shown in the bottom portion of' this

It is possible that the terms "solar particle flux" and "solar figure. Table 2 shows that significant solar proton events oc-particle fluence" may be confused. Particle physicists usually cur in episodes with a large variance in the distribution. Thererefer to the peak flux observed in a specific channel of a solar can be relatively long periods between significant events dur-particle detector. This can be either an integral flux above a ing the sunspot solar maximum; conversely, significant solarspecified energy level, in units of particles (cm 2-s - I-sr - ) or proton events, including ground-level events, have occurreda differential measurement that specifies the flux at a specific during solar minimum. Of the 201 significant particle eventsenergy in units of particles (cm 2-s -sr '-MeV 1). Individ- during the past three solar cycles, 34 of them have beenual events are usually compared using identical channels. Peak ground-level events detected by neutron monitors indicative of'flux is usually used to describe solar particle events. Fluenee is solar protons with energies greater than 450 MeV.the total number of particles above a selected energy that areexperieniced throughout an entire event. Fluence may be given tData Artifactsin either directional units of particles (cm 2 -sr - 1) or omni- A ca:eful examination of the data-acquisition techniquesdirectional units of particles tecm 2). The fluence is generally reveals a number of artifacts that may obscure any long-termof concern for the total radiation exposure. During an episode trends. The identification of solar proton events by interpret-of activity, there may be a number of individual solar particle ing ionospheric data is affected by at least two pronouncedevents that contribute to the total radiation exposure. Under seasonal effects. The rionmer is most sensitive to the sunlitthese circumstances, instead of trying to derive the fluence per polar ionosphere (its nighttime response is about one order ofsolar flare injection, which is extremely difficult, a pragmatic magnitude less sensitive), and since most polar ionosphere ob-solutton ts to obtatn the fluence per eptsode of activity. servtng sites are in the northern hemisphere (parttcularly dur-

In an attempt to investigate the solar cycle effects on solar ing the 19th solar cycle), there is a distinct northern hems-particle events, we have assembled a list of solar proton events phere bias in the ionospheric data with more events reported inthat is as homogeneous as possible. Our criterion for this event the northern hemisphere summer than the northernlist was to identify all > 10 MeV solar proton events having a hemisphere winter. The ionosphere also has a strong responseminimum flux of 10 protons (cm 2 -s'-sr 1); a criterion that to geomagnetic activity, which has a statistically significatcould be applii;" wi;6omly over three solar cy :!.- Q--,craft- peak during the equinox. These two effects combine to give ameasured proton fluxes were used whenever they were very tionuniform temporal distribution of the data assembly.available. For the pre-spacecraft era or when spacecraft data particularly during the 19th solar cycle as illustrated in Ftg I1.were not available, an equivalent proton-induced polar cap The magnitude of the seasonal bias is not generally appre-absorption event was used. A practical "rule of thumb" ciated and has crept into data sets considered quite reliable.useful for converting sunlit polar cap riometer absorption to Satellite-sensed data assemblies"4 still show a seasonal bias atproton flux is J= bA 2 , whereiJis the flux of protons with en- the lowest flux threshold as shown in the top part of Fig. 12,ergy > 10 MeV in units of (cm 2 -s-sr) Iand A is the 30-MI-l which displays the number of > 10 MeV solar proton eventspolar cap riometer absorption in decibels, Thus, 10 protons with a flux greater t han I part icle (cm: -s-sr) '. I n 1 his case, we(cm 2 -s-sr) Iwith energies > 10 MeV are approximately believe that the apparent maximum around the equinox is real,

NOV-DEC. 1989 PROTON EVENTS DURING THE PAST THREE SOLAR (_ Y(I.ES 409

Table 2 F> 10 MeN solar proton events with flux > 10 (cmz-s-sr)

Year Ian. (eb. Mar. April Ma' June July Aug. Sept. Oct. No%.. Dec. Total

1955 I .

1956 I I 1 4

(957 1 2 1 1 2 3 4 i 1 161958 I 2 1 2 4 I 11

1959 1 I 3 17

196(1 1 4 3 14

1961 4 8

1962 1 11963 1 2 31964 01

1 65 I.

1966 1I 51967 I . 2 .

1968 1 2 2 2 3 2 121969 1 4 1 1 2 I (3

1970 I 2 I 1 1 81971 I I I .1972 t 1 91973 l 1 2

j974 4 31975 22t976 11977 3419 7 8 1 . 1 2 . 1 3 12

1979 1! 111980. 1 12

1981 . 2 2 2 1 2 I11982 1 2 2 2 4 121983 1 2

1984 2 2 1 2 71985 1 31986 I 31987 0

Jan. Ieb. Mar. April May June July Aug. Sept. Oct. No'.. Dec. 1 otal9 21) (2 2J 18 (2 29 17 29 6 (8 10 201

In Solar Proton Events and the Solar CycleL )t : ' /'- '-" - -'In trying to determine solar cycle similarities or difference,

in solar proton events from the past three solar cycles, ' e usedthe monthly mean sunspot number". as our major ordering

paaee We used month of the minimum in the smoothedi -1 sunspot number to identify the change from one solar cycle to

2 the next, since the monthly mean numbers have a Avide %ari-ance during solar minimum. The top portion of Fig. 13 show s

-, the number of solar proton events that occurred each 12-month period after sunspot minimum for the past three solarcycles; the 12-month mean sunspot number of the same period

Sis shown in the bottom of the figure. A summary of theRo 'number of events for each solar cycle is given in Fable 3. The

sthree histograms in the top part of Fig. 13 are all different,2 ;--I and beyond the obvious fact that there are more solar proton

4- ,. events during solar activity maximum than at solar minimum.we cannot discern any repeatable pattern other than the gen-

Sral association with solar activit. The only constancy is thatthere is an average of six significant solar proton events per

YA year independent of the length of each cycle. Using solar mini-ig. 10 Discrete solar proton events per year (top) and the yearly mum as our fiducial mark, we have added the values for each

sunspot number (bottom) cycle; the results are displayed in Fig. 14. The data are or-ganized in 12-month periods beginning with the month afterstatistically smooth sunspot minimum. From Fig. 14 it ap-

possibly reflecting interplanetary acceleration processes. pears that the majority of solar proton events will occur fromAnotmer data artifact exists in the 19th solar cycle proton Flux the second through eighth years after sunspot minimum. Theand fluence data as a restlt of exponential forms used to apparent peak in the number of proton events in the 10th yearmodel the data'" " "' which introduces apparent systematic after sunspo! minimum is an artifact resulting from the epi-behavior into some events, particularly the events derived sodes of activity in the 10th year of solar cycle 20, which wasfrom interpretation of ionospheric measirements. When the 14 months longer than the other two cycles.extramagnetospheric satellite-sensed solar proton event datafor solar cycle 20 (such as NSSDC data set 69-053A-07("') are Episodes of Actisit,summed over event intervals, similar systematic patterns are In compiling the list of significant solar proton events, wenot found. tried to identify each event with a solar flare on the sun. In

410 1). U. SMART ANt) M. A. SHEA J. SPA(Ci'RAI.T

Y F many cases, there %%ere multiple flares on the sun, all of ',hichma, have released particle, associated "ith the aggregate par-ticle event observed at the Earth. There Aere two type, of ,e-quences of activity, the most common being multiple particleevents associated with multiple flares from the same active re-gion. The other type of activity sequence occurs when dif-ferent regions on the sun each produce copious solar particles.We hase calculated the number of discrete solar proton produ-

-Y. -,ing rei-ns associated with proton even:, detected at theEarth for each of the last three solar cycles (i.e., multipleeents from the same region contributed to only one episode).These results, listed in Table 3, show that for each of the lastthree solar cycles at least 17%O of the significant solar proton

- events observed at the Earth are from solar regions that pro-duce at least two or more discrete proton events.

;z 5 Stat ieql I imitationsi

The small number of esents during the past 33 years seserel,limits statistical analyses. Typical spacecraft engineering pro-cedures ask for reliability analyses \Nith 90"o confidence lac-tors; Table 2 clearly shoss that there is not enough data to sat-isfy such a requirement. The practice of dividing the asailable

F-- - data into solar cycle groups further limits the statistics, andthe results are open to a ,ariety of interpretations that cannotbe squelched by application of statistical techniques.

-- _ Despite its limitations, this data base or a portion of it hasbeen analyzed in a number of different ways to develop statis-tics and models of solar particle event frequency, magnitude,and fluence." . -' "' The 1975 NASA model" is still in use to-day and is the contemporary standard against which othercL wsork is compared. In the past few years, this model has been

3 .... 9 attacked as being too limited and not trul, representative ofMCNTri

hig. 1I iNonuniform temporal distribution of proton events for threesolar cscles.

SOLAR FLARE PARTICLES " K =- ' -1963-1982- I-J,] .i.Kl1' r - .. .

IY'_E '9 :Y: [ 2' ,Q..E J.pARTIICL9 FLUI) GRIA14ll Y-N

19.. I- b-S14 -__ ,40- , ,

12 :.-

Solar cyle

100

z

2I 17H 1l i t2 vii f

No o of integrate

JAN FEB MAR APR MAY JJU L AUG SEP OCT NOV DEC YEARS rROM SiNSPOT Mi4iM M

MONTH OF ONSET Fig. 13 Sinifica l discrete solar proton events for each 12-monthFig. 12 Illustration of the seasonal bias in solar proton data. period after solar minimum.

Table 3 Solar proton events for solar cycles 19-21Solar cycle

No. of No. of integratedNo. of discrete discrete solar proton fluencemonths proton producing

Cycle Start' End in cycle events regions > 10 MeV > 30 Mev19 May 1954 Oct. 1964 126 65 47 6 .7 x l1i 1.1 x lol

20 Nov. 1964 June 1976 .41j 73 56 2.5 x 10" ° 7.0 x 10"

21 July 1976 Sept. 1986 123 63 52 1.8 x I0l' 6.1- I) '

aThe start of each v, 'i cycle was selected as the month after the minimum in the smoothed sunspot number.

PROltt~ l~ I(N [I N I'S W RIV1x li-i11 A5 sIIH R VE Sol A\R C) ( I I S 411

30la c ,rii cnsad cs c'rr tliicics. I tic 20th solar sc30 sas thle tuisl solar actlll is\tci c that i mas \stenrti ailsa oh

sen ed ursingi '.ace-borne inttsttniciatiirn. I hie data base torthle 21 qt solar ci '\Ic cott~irs mil.\n sinall Ciel etif tat base been

W Z N w ~ell obsers ed b\ contemtporar\ scttiitiie particle detect litI> 20, hoseser. the ci en! integrated fluenice of these "sniall ci ents

7:) UJ is %cr\ sinall comipared to the major es ent - Thle analk is hZ I Clntriati et a!. Contclurdes thlat all Of the fluerice data for the

joI three solar cicJes cottthitie ito forrtr a conitTUOus ILoirtiortualorQ I) distributio that is represeiinrtt-'e ot ,olar partic hisor\. Itt

U 10- Icontrast, otheri researcherN, " Lsi ti smrallecr data sets. f oundI a srmaller distributrioni funictiont or is hich the \er\ large eci it

0 Iare not1 part olf the -expected" log!-rorttial distrih ution. and sohe>c calculated intdepenideritlN. the proa hi lit ot a s en> large

L LI- able 3 presents thle solar c\ c suirmned fluiices %is i i criergies greater than 10 and 3t0 Me \ Th cart be iterpret ed as

160shoss in a ,vNternatic don tnard t rend: lionsien.r tins trendI6Or- ~does lilt rmatch thle mramrrititi suntspot riittrhet fortile rcspc-

iA 4Q _ tiSc c> des. 1 urthIertirore. tile di ftet etce bet isccii 20cl tintt(D C21 integrated flretices Cart he ttterpt ctcd a, lie occ:urrenc (or

C 20- tnornoccurrertce tnear tire I-arthl of one \cii at ge 'l1ar particleW episode.

Z M ooi-Organization of the Solar Particle IData Rase0 eo- it Other Parameters0 60- (hs iolN!>, solar proton ce tt data can be orgarri/ed aga inst

60tmil\, solar parameters such as, Sunspot rnunmber, I f-ctin radioemission, solar flares, solar x-ras s, and so forth. There is no

- 40-- reason, at this time, to select an, particular parameter os erI' art other. We has c compared tilie number of sigri hcatrl solar

2 0 proton esents with solar tlares of importance I or greater.'~I ~ ~HoNs Cir, Cs Ci for solar flare data mart> cas eats, must be ap-~' -perided because during the past three solar cycles, there hase

been drastic changes in observing sites, reporting criteria, and1 510 data grouping specifications. The solar hlare da~ta shown in

Fig. I5 wAere derised from N(iI)Ci0l f Although there are gen-YEARS FROM SUNSPOT MINIMUM eral trends, betwseen pr oton e- ents and solar actis iry, as identi-

fig. 14 ',ummrinion ofi simnificant discrete solar proton esents for fied hi, either thle Sunspot number or solar hlares, there is nodtvle 19 21. ohs ioins linear relationship. Hoss e r. the -\sariatice'' in the

niurmber of sigrrifi carnk Proton esents per ,ear is consistent is iththre range expected fronm a biromial distribution or othermathetratical techniques appropriate for the statistics of small

ihit atn o),crir ( uriouslx there are tiso attacks,: 1) that its numrbers.prediitoti arc too high because s er> large fluence es Cuts sere Fhe IFebruaryr 23. 1956, solar particle event is thle classic ex-not obsers cd hi\ Iartil-orbiting satellites, during the 21st solar ample of a discrete high-energy particle injection fromn the sun

Ic.atd 2) that it does riot properl> include the posibilit> at a "'avorable'' propagation position iwith res;pect to theth en. large fItetc ci cnit' could, and hrstortcall>, base oc- Earth. -The solar flare that was, the particle source occurred at

reck ~0334 Universal Time hUT), sometimes referred to as Green-wAich kMean Time, at heliographic coordinates N23, W80. Pro-torts with energies greater than 16 GeV arrised at the Earth at

Worst-('ase Models (1345 1 :E I mitt). The energy content of this solar particle eventi tes eloping is orit-case models, the ,er> large or icr> ell- has not beeti duplicated by Earth-based measurements in the

crnreti, particle vseirts are ctiployed initead of the common three solar cycles since the event occurred. Since this event waseserlts,. It is orur opinion that the sery large particle events may prior to the WGY and before the "space'' era, ground-basedbe expected to base "ritial'' elemiental abundance ratios,. measurements werte fragmentary, and some of the early ana-F or this paper, ise "ill suggest two solar particle events. for use lyses are deficient wihetn viewed with modern knowledge. Theats extrerne-case mordels. [ he first isp is a "'classical'" serv ern- high-energy flux estimate "was - I50 particles (cml-s-sr)'erget ic sitar particle event hasting at discrete ijection of flare- at energies > 450 MleV. [his was riot air exireniiely largeaccelerated particles along the idealized interplanetary propa- fluence evenrt at etrergies of about 30 NMeV." " '- Interplane-gatiorn path from the sour to the Earth. For this example. %c tary conditions were reltisely ''quiet'' %%hen this event oc-%sill use the Februars 23. 1956, solar particle event. The second curred. arid the particle flux decayed itt a classical mtanner.iype' ol esenit is the injection of a large population of'solar par- Alter about 12 h, the very energetic particles had diffusedtides that is reaccelerated by interplaneiary phenomena hefore beyond the Earth, and the rentaining particle flux could bedetection at the IFartlr. F-or this examnple, sie shall use the treated as an ordinary particle event. However, during the in-August 4. 1972,. solar particle esent. itial 12 h, there was certainly a large, ,ery energetic particle ra-

T-he fluence data hase countairns a dletectiur technique ar-tihact that essentialls cannot he remmnsed. [he tremendous ad-%ances in detector tchnrology are such that satellite sensors ',Solar Ilares aire classit ed hs the irei; fI tie hlare on a scale of0 tronow rioutinely detect esenis that were far below the detection 4 1 lire sitiallesi tj bflares ithiose %sith at rre,, d area less than 2.06threshold even I5 years ago. [ lie 19th solar cycle is deficient in stirare deg) are ot importance ii \&hctreas extreme> large flares Ithose'.small event'' identificatin; horwever. it contains the largest isith corrected areas greater titan 24:7 square deg are of iniportane 4.

1) 1ii SMl*\R I \\M l. VStll-A I 'I'M( R <l I

p> 10 NeV15--Iflux > 10 cin sec 1 ster'

I.J

:j7

1960 1970 1980

77-1

V77

L~100-

o - II

960 1970 1980

YEAR

Sig. 15 % earl% nurmber of solar proton rsents flop) and solar flares Aijth importance >I

d iation exposure that si ond a fftcct %u IunerablIe devices atnd con- gradients. In contrast, howkexer. onl AnuLs! 4 at larecr ftix is,niea radiation ha/ard. obsers ed at thie [ 'arthI than at the P~ioneer 9 spacecral

[he solar paiticlc esenit most ottcn used as a sot-ae It is our Opinion that the Auguist 4, 1972, solar particle flu\itiodel* is the AAtreust 1972 eiiodeC of' solar acti its because pr otile observed at the Earth reflects a sequcelC of unique11 andI art h-orhvi~ing spacecraft nleasuremetts of the particle flux uinusual occurrences: the result of' a large in ject on of' solar-iud ltluetk exist. The solar active revion t sas located at east- particles into a region of- space "~here the cons ergitie in-crri heliographic longitude: frotm thle aspect of' the Earth this terplanetary shock structures reaccclerated wshat w~as a mub-%%ias an eastern hem sphecre cx ent sequence. During this time, stant ral solar particle population into anl extraordinar\ solarlie Pioneer 9 spacecralft "sas at 0.77 a.u. -, 46' east of the sun- particle populatiott. (See I ee' "' for a more detai led discuis-

I arthi line: fromn the asoect of the Pioneer 9 spacecra ft, this %ion ol' shock acceleration.) Early onl A ugust 4, just prior it),s as a wcestcrri hetnisphiere event sequecnce. Fihe initial particles, the flare, two geomragnectic suiddctl coinmnetcenien ts s ere rec-hat Ac er obsers ed s ere getierated by a solar flare onl A4.gust 2 orded at the Larth Ii.ndicatix e of" thIe passage oft solar-

at 116 U LI at a heliOg-aplhiC lougit tide 34' east of the sun- generated shock wxax.es from flare,, in thle solar actixe: region[arth hihne. (ie major particle et cutf that was observed at thle that produced thle major en ts. I Ilte shock structn rcrier -

[arthI onl Atiust 4 \xas generatcd by a solar flare at 06201 UTI, ated by thle Auigmst 2 solar flares had alreadv propagated9~ e1ast 10 tilie sun -farthI line. The solar proton tinme-intensity bi-sond Piotneer 9 onl August 4 when t he mrajor solar pat t iceltistor. ol early August 1972 is s.hosx till 1ig. 16. The Pioneer 9 injection occurred.) When th lhare of August 4 occirred atldata are for i _ 14 MeV protons indicated by thre hecavy line. 0620) UT, tie initial interplanetart, shrocks, had jutrs passed thleI lie Earth-orbiting INI P spacecraft data from protons with en- E;,arth, lcaviiig the Earth etixloped bet nxeeri the first shock errergies > It0 and > 30 MvcV are indicated by the tim line, Be- semble arnd the much more posxert il shock generated h\ theginrining AxithI thle solar flares, from the same active region on Arugust 4 flare. While t lie I-arth \i xas ens eloped bet\.c tfeei resAuList 2 until Auigust 4, the particle flux observed by thc Pin- two powAerfutl converging shocks. the flux obsers ed at( thericer 9 spacecraft wias larger than the flux observed at the Earth vnas higher than t hat observed by Pioneer 9 Thle titeEarth, as, would be expected fromn coronal propagatin and period when the Earth wxas betwxeen thle converging purw&ert nil

NOV.-DEC. 1989 PROTON EVENTS DURING THE PAST THREE SOLAR CYCLES 413

iO F - Table 4 Comparison of the first 27 months of solar C)rcles 21 and 22PIONEER 9 38 Cyclc 21 Cycle 2C 77AU, E09Cl - 2

104 ESP = 46'E PIroton events I0 9P > 14 MeV x-ray c.cini, (s-level) 19 13

Smoothed sunspot nUtmb er' 64 S4

to , "I w nionh 20 aMwv 11oIa1 imhiuun.

* " >10 MeV was similar to that of August 1972. In July 1959, a series ofCo~f ---.---4 " .\"-\ 1 solar flares near the central meridian of the sun produced ma-

/I jorgeomagneticeffects with resultinglarge variations in theI \ >30 MeV0 10 alactic cosmic-rav intlensity as measured by neutron monitors

that recorded several "Forbush" effects in addition to a small"flare-associated" relativistic solar particle increase. These

00 observations were very similar to neutron monitor obsera-v v tions in August 1972. We believe tha if satellite measurement

had bcen available in July 1959, the tluence from this activityI YL 2 episode would have been equal to or greater than the fluence

2 3 4 5 6 7 observed in August 1972. In fact, the fluence deduced for [heAUSUST 1972 July 1959 solar activity episode exceeds the measured fluence

obsrve fo th Auust197 soar ctiilvepisode 4 4 v2Fig. 16 Solar proton time-intensity profile oserved for the August observed for the August 1972 solar activity1972 sequence of e,,ents. It is our opinion that the extraordinary events, such as those

that occurred in July 1959 and August 1972, are the result of asequence of major flares, particle events, and solar-generated

interplanetary shocks is the only time when there is anything interplanetary shock structures which contribute to the unusu-extraordinary about the observed particle flux. During this ally large effect. These events are outstanding examples. Thetime, the Earth-observed particle flux was unusually high and fact that the Earth did not record such a flux during the 21sthad an extraordinarily hard spectrum. This time period, from solar cycle does not mean that these type of events did not oc-about 06 UT to about 24 UT on August 4, is illustrated by the cur somewhere on the sun. Indeed, we wish to use an analogyshaded portion of Fig. 16. After the converging interplanetary from terrestrial weather. A hurricane or a typhoon is the resultshock structure had passed the Earth, the Earth-observed of an unusual sequence of events that are likely to occur onlytime-intensity profiles resumed a classical appearance and during a certain season. These storms come to public attentionmatch very well the results expected from application of pro- only when they strike populated areas. Likewise, extraordi-ton prediction models such as those of Smart and Shea"'' nary solar-initiated storms only come to our attention whenThese observations also suggest that these unusually large flux they envelope our planet. Our sensor net in space is so sparseesents can be limited in time and spatial extent, that we are not likely to find such events elsewhere.

In using the August 4, 1972, event to study major particlefluxes and worst-case scenarios, the question often asked is Solar Particle Events in Solar Cycle 22"'Should an adjustment be made from the observed flux and Two years of solar cycle 22 have passed and although the in-fluence to a w\,rst-case model by invoking coronal gradients?" itial rise of the sunspot number was the most rapid to date, theas discussed previously in this paper. (The argument for doing rate of rise now appears to be slowing. The number of solarthis is that the solar flare did not occur at the most favorable proton events observed at this stage of the solar cycle (apply-propagation location for measurements at the Earth, and per- ing our discrimination threshold of events with energies > 10haps if thi., flare had been on the western hemisphere of the particles (cm2-s-sr) I is within expectations. However, thesun at a heliographic longitude of about 600 an even larger first relativistic solar proton event of this cycle, comparable toflux would have been observed at the Earth). We argue against February 23, 1956, July 7, 1966, or November 22, 1977, hasmaking such an adjustment for the August 4, 1972, fluxes ob- not been dete.ted at the Earth.' Recognizing that we are deal-served at the Earth and suggest that this is an example of in- ing with small numbers, the statistics for the first 27 monthsterplanetary acceleration modifying the "initial injected" of this cycle compared with the 21st cycle are as shown inpopulation of solar partcles. A comparison with the particle Table 4.flux measured by the Pioneer 9 spacecraft definitely does not Solar maximum is expected around the first part of 1990,support the "flux adjustment" hpothesis. The Pioneer 9 data Perhaps some of the paricle events that we think should haveon August 4 have 1'een viewed with some skepticism precisely occurred by now will occur during solar maximum. Perhapsbecause the flux measured on Pioneer 9 is not what would be 1990 will be like 1980, with only two significant solar protonexpected from the relative positions of the two measurement events at the Earth. At the present time, with the sunspotlocations with respect to the flare on August 4. However, the number up and the number of x-ray events down, it would bePioneer data are considered valid for scientific analysis before foolhardy to try to predict the number and/or magnitudethe August 1972 events and are again considered proper for of proton events for solar cycle 22. All of which goes toscientific analyses after the August 1972 events. We suggest prove-there is alviys something new under the sun!that the Pioneer 9 data are also valid during the August 1972events. Thus, the "worst case" may well be what was actually Concluding Remarksmeasured at the Earth. To support this argument further, we This paper does not have a definitive incontestable conclu-note that Lingenlher and Hudson 2 have argued that the sion. We have tried to present the solar particle data availableanalysis of returned lunar samples and cosmogenic iso- from three solar cycles in an unbiased manner and point outtopes indicates that events from our sun with fluence greater the limitations, deficiencies, and pitfalls inherent in an analy-than 101' protons (cm 2) with energies > 10 MeV are veryrare. In addition, (ioswami et al.' found that the mean fluxvalues for the last three solar cycles agreed fairly well with the Ithe first ground-lecel 00ccl of the 22nd solar cycle occurred onvalue deduced from solar flare proton and alpha particle in- August 16, 1989. [he 5.5-Year interval hetween the ground-levelduced radioactivity in lunar samples. ceents of I-cbruary 16, 1984, and August 16, 1989, is the longe-st time

There was at least one other period during the past three mnueral hctcen ground-lel c'eents %ni1cc routine monitoring begansolar cycles when the particle flux from a sequence of activity in 1953.

414 1F). I . SN IA RI A\NIJ NM. A. I-) I A i. IsF'KV(I ( R \1 I

,i, kit this data base. [hfe iolence of thle 19th solar cycle ssas u lre. ora I pt ua R's.,ar, it. V ol, 9m), tlsk l9i.not totallsI repeated tin thte 20th o'r the 21 st so lar k.'cle. I here pp. 62CS oi200.Mie xertain characteristics of thle obsersed proto:' esent distr- "\I,( raii , Is o ITO No. f I -s< ~ (14 'sill. khLItiOnl foun1d h\ alt inls si1ators: thle distribution of evenits is 11he/Si,,' '5, h'it,t' Reiw- \,, 11N ~. (), N-0~i, pp

descibed Is \ 41-ntormal st'itist ics. [hle data base is limited, 211

anldepisi-des olacrisits domninat e thle occurrences of' solar par- "ML .IcttlliK, h Slur Ir IZ F cUat. k5 P/t ll K t. I IP-

id:e cln ms. 'roba hilits distribution oft expected occurrences 'k at 19" , l. I't III I I'ittl ha\ C beenI, anld \\ ill c:ont inue to be dcris ed. Ft is not clear ,, Oot N .R- OeI1si. FI ',,)It, I t . and Isneect. \ s''-

tat there is a ss sterratic sola: crete related beltaizor inl the lationt ot F .it-cac(otiai \i Sn lLI,(III tild ( 'stttu kNissdata. 'Such trend' ni\ he prescrtt hioseer, the nonluniF'ormitv SoLur \'iimd anid ( r,iuil liiilutice ,t on a rIts De:rclisc F im(,t [tie dat and tite dominance otl episodic sequecnces, of actis- OnTe Solar Roiait,". l'rme-ii'nes -I/ thte 14th Intiaftwnt (iotIt\ ohscnurc ait ohsiotn trenid. [hle as ailable models mia\ be Rat Conjvernc' 'lfun,/t .VI 1, M\ P-Flaici t'init tit F \nitte

ai e s -coniseratise.- such us the IFevumian et al." model reitci 't ktiitcli. I RI'. I' , pp. li09s 1099,

based on at analssis of all as ailable data, ''aierage.'' such as - R oclot . I , C ''\,:\% \ spC,-I 0!1 Illtte1tIIA I 1 t'i wa i w.iti Rtie NASA itodceV based ouR ott direct satellite obsersatiotts, scaikd h% 01.3 Nlc% NO',Ilt P101011i F \C1111r I16-.'' Setar I,'rre'smra/

Rlum,m . I. its\ o ( iler I vai alixu . (iidi. 19-1, pp -41 1 4t 1or Iliheal'' 'sich as thle 'hentcte>'' model based only ort the - olt . knF~teeFti~si, ~ u'I',j 'atnost Itnoi' rcent~l solar cs cle. V-herct'ore. any judgments-I, of' the (an lt,in't. I olAA I..-\,inci I Ie'lsl I t, .n Sit ,t'aitt

icepah i t a speciftic nmodel should be based onl the risk Ion. M)(, F197f, pp 214 31illt (o6I ed. , 'Dlodwin F'itt~. IF ,t . Foc,inui FI ,,tmid \Fo'lc . 0) ID..

-Sttrses antd ( Tttttaris'i o1 ,lx 51 iktSiti F Ie ct t:tilIFliitssoit i I ). 19\0, V i Ie (s't'(slist I .t, , FFrttts , -Nt Ft. \t -N

Refe'rences A *F(it I R -'-0222, Sept. 19'>N~ii't .I . ''si ( sle Iileci, ont F rapped I ereic P'ar - "I1 itt, R. F'P , 'oltt F',,ttdc V:,clt,ii,twn antd I'tipl..

,c,-~ /'irui .1 'ta,'airatt,l Rt ts N01,l 26, No. 6, 1989, ppt. Rciew (, Geop I (;'~ st, ~. N oI 25. \IpT 11 191,. pp. 6-6 (684.1P, -' Nt i'it. 0 . \t.I lit- ( olltptstt'Mo ' u il i ( 'utlc Fa' antd

)kan, 1, 11I . Ir., 'siliberere I.. anid Itaso. C.H , ''Cosmic Ray' Solair F ntereeic F'aittcles''A, ' u 'i (,c'iw- Wt \,,N I. pI Nt~t l \1i)CICevitis F'at tL ttle Near-Lartlt 1article 1:nsiroin- 198- ppt. 68 696'

11-1t, N is~tl Resecit~t F Ab., Vtastirtii.n DC,' NRI Nieno Rept. D_ 111i1, N.F'.I<N.Foeiti.t i (,I'ccklct. (-ii(~N '.19.s. "A SWIis OI Nte %1\ UitCl'i 'Oldit I 1,11C 1It~~iiC V, 26''pAlt 1 ) 1 .itld Ylha. Nt. NV. "Galactic ( osmic FRadiatioit aitd l)ttt I.U tlte 19- - 19'" S,'L~ir \1iti1iillittt1 F'et 10d, I s troph I/tt s11 ./tir-

rio h '1fc ittce llantl/',t oI/ (it,ophiymt' and fte S'tptlL nal, 1 1. 239) Augv. 190 pp. lii-IN/ ,> 'l'PS',l I .'tie V' N. s T11 .lr.it Nr orce (Aoph'.sics tab.. FBed- (,ttcerct G. -( totiiptstil tIt teteIII~ cti. F'a t t~c F'utlalilo. Ill

\1 NtN\ . 1 (Itll 6, Fitterplatitar% spuic' Rt w, S' ol(,ph i ' (d'l. I I title 1 9'9.1c ones. I I .tpsl Wstrd. % k . 1-"(osii Rays it tlie Fuieurplane- lip.

569-58t -5 .tedviil''F te,',,t 1. \/tril Nol. 14', Nictt 196". pp. Cook. W\. R.. Slone. I.(. aitd \N''etR Is " I lenientat ( ottpoi

Ii f I Is i sott oi' Solar I-ntergei e Iar ticfes,- F strn p/i ts ttl .imrnal. Vol.29(Iit\Ituri.'. \1. t,tsaon. (, \1 . atnd SimtpsonT.... lite .(tit\ 1984, pp. 82 >838.i,~l i I( omtpontlett ill lie ( ostiic-ras Spectruni at -5i0 -"'McGuire, Rs. LI. sonl Rosei initz, FI I . and Mce~onald, I . FB..

Nt pe NiC \1IC' i IttiiIT' 2 P)-4,- -F s riphisital .Ioiirnal. Vol. 'F he It 'outposition of Solar L ucreitic [)articles,. F trop/i vs ipur-2' '. N.. eIIl'ei I 9-. pp. 2h; 2- tol, V ol. 31, 1 ch. 1986, lip. 938 -96 1.

*~ ~ lie itlilic . \L F I\ I 'teS~ t it ICt: Flitors antd 'rcdic- "Tmel5.1 FT..te I.ntergertce ot Radio Astrontuns inl tie L'. K.bawwit' i ,f t 'tjstura antd R i1'Afits, N ot 26, 1989 -Nti Cr World W ar IF,- Quarterli,n orial o/ f/it Rot ci F Irontoft, a

NVlatn,. I FtIT 'h ottic-Ras FI flects oit Nlicroctectrortics, .Socief'. Vol. 28. March 1987, pp. t -9.I N.'\.,tI Rcsearch I dv., Vkastiitt'. MI , NRI Miemo Rept. ti Iorbst~it S. L:...tIlree L nIusulal ('s'sttic- Ras Futensits tI creases

'901. FDc. . ) 6 Dus' )te to Chlatrged Particles Fro rt ie t Stitl,' P/i vuwal Ret ievA, Vo(l. 7(),

- NF.Kiht'et. ks. Ft .t,latci, ( ostit Ras aitd Aniomnaltots (.oiii No%. 1946, pp. 17177 .2.p( iti ie hc i'soplicc. '' Rctt'ws 0f (;iop/iisics ,Vol. 25. April 'I'orbhtit S. L., Stintteonib, 1. F).. attd Sceli, NI.. *'File F-

19 T' I I 2 - traudinars' FIterease of Cosmic-Ras F ttmensits onl Nos ember 19, 1949,''Sslu es t lu Na /rts. I o'op/i it's and( Awr (op/i c Ms'sAonto- Ph r.mcal Ret 'cc, N ol 79, -Ally. t1951), pp.1 tO -. 014.

cruph/t' N,,I 5.Redcl. torslrechlt Holland. 19-6. "Simpson, 1.t A .. '('o,ic-Raditioi Neutron FIntensits Nlotlt-'' us ci. F\% . Kaihler. ',. V .. ( uri, H1. V'.. Kotiteti, NIo. T o .Atnals o1 fth'e 11, VOL 4, F'ergamtori. F ondori, 1957. pp.

Ntvluc,. 1) 1 atidl1hxN Suds. . J.,''Ihe 01i I'-Associated Flare 3,SF-373.21 'tie., I. \1h''~la 's titVo. 89, Not. 1983, lp.v 181 193. 4('armicliael. H..."(osmie Rass Flmtsrmerttsl. Anuas o f fte

Yhea. NtI- atid 'Start . 1). F-. Retati'. ste Solar Particle Ls ettts IQSV. V'ol. 1, N1I' Press, ( armbridite, MA. 1908. tsp. t1'8-1F97.'t iri t -hs' \1.\ ',\N "' 'So/r Alanmmrn '-F ,alrsi.s. edited bs V. F. "I t~FitlI, C'. 6. antd I eitnhaclt H.. ''ite Risouteter-A Des ice fs'rS1icpiti's MITIlN N 's t)hrtudkt' 'NN press. I itrccht, the Neterlantds the C'ontinuoits Mcliueremtl tif Fomttsplteii ..\horpmiomt,'' Prot'e'ed-I 9s- pP, Ilt) 11 4. Int's /- I/it' Ins ft iti' 11 Radio I:iten,s V'oh. 47(2). F1 eb - 1959, pp.

- istltet. 's %t , ( litr I %kt ( itte. [I. V..N(iuire, R. F., 115-12W)ksre (, , anid Steele's. N. I.. It.. -Solar Filanterm LIrupltis arid Ntalitson, IF. HF. F Tables to 'Slairt'ron Iketits,'' .'tt I'rion

F ixervcrt, F',itte feti, Fsfrt/i/tisii'a .htrna/. Vol .3012, Mardi Afaltua/, edited 6% 1' . cF~ittalut. NASA, \\ Vasiittioii, W)1, NASA1916. pp 104 1 li 'I R-F109, Se t . 196 3, pt, 109 - I1I

1 i'tcilt. I P' . 'Nrits, I., atid Mitchiael, 6. A-. ''-InI FDetisits BIailey , 1). K,' I'il at ( ',it p 'bhorp't iont," ' ' Iliatud .'fat't 'tel.Nt..ttITCtTt I Ot I'fS Rdji FHTIrsi otf Protonr- PrtoductiFig I lares and cre, V'ol. 12, Mar 1964, pp 4t)S 5319.Nttiputilt I kiiesj' .Imirna/ o f /eoitt-u/ Researt'/, Votl. 72, No\ . "\art ItioltehekL. \1 VN I.,M ' til tig. I al.,id NMeF~iirald, F . FB..196-. pp '4') I 49S 'I tie Variatiott oh Solar 'tooi I nrns 'Stteetra ,Ii 'Sire[I Ittitiioli

'astell. 1 11 F' tist (mitstte, 1). -V. ''Ittpat of ( urreni Soitr stit I lehltltiitntitle,"' 'tiar /'/iits. Not. .II, Marchi 1975, pp,Rado ';iro (Fses~itttt, u, mrsfrontom,tt.Vol. I'),edited h 18') 223.

A. Fleer anid F' [feer, F'ergviiii, I oriditi, 1976. pp,.355. 384. "'iSsesik,i M..rITI SIttii, P' (Cds. . ( ',tld'' itI 'io/r 'titt I I(-I[rf.-I iselit. 1 11. int]la ,tttsriiti. (. I ., ''A. ( 'aiattg (it Prttiir PP.5 / 909. AllIttthisICs Iid 'spAie SICICIICC F Ihr\ . i N. 49. Reidel ,

F senil. 1966' 19-6 Filsiic Nollt( laSSIcal Soilar Ractit Spectra,'' Air DIitrrcltt, Hoillaitd. t)"5

F 1s 1~ (eiphissics I ,th , IFatiscii AI It. NIX, AT -lu -! R78-0121, '''-'kittis att .) 1.., tta/itesskasa. (, -N, Ilo, V. N'Slat 197IX %iro,siichlitkii I , Ft.. Nsit/iiisai N\1 N , I'ctes~rslita, . Is , Poigo-

i"KiiI,Iet. 's V I hie Role tud tic' Big F-tare 'sstidrnit mll ( 'itrela- dIrt. I. I . 'sIdidktts. \N I.Ile\s, V' -V. aintd ( trltik F. MI..('tatoetitis ~o Solar [trieriteit F'riw,it anid ANssociatedF \Iicriimase Parartic )/ 'tolar /'ufii /-tin/. /'0 i /9-v. Iti Ru-itsit) 'Ncaderim tit ',,itrts," ' oirnta/ t/('o -sli R,',earth. Voil. 87, %SIas' I 992. pp tcii.- s tilte I'sR H.5.i CII III I (i tll (111 'itsolalt Ier rest rat Ir scs1419 3408 'liiCOcc. I1982.

It' her, F-. \ki - McNamara, I . I .. and (jetildc, I " ,'Peak-lbsu .'ti/izr (iiti/t/iiu t-al Oulut, .itt,'tTal I ip,cii~Tlt 1)111 ( eriel,IDersilr Spktra oif I ttry'' Solar Radio thirsts arud Prtoni I' missioni Bouiilder, ( 0

£ NO\'.-[)I:. 19K9 P'RO ION EVENVS 1)URIN6.THE PAST THREE SOLAR CYCLES 415

- - shca, mI. vN ait Smaill, . 1I Sgttcn Solar Protonl FesCt, sion',,' Journal of Spacecrafi and Rockets. Vol. 11, June 1974, pp.S9 5 -l969.' So/ar. lerrestrial Phisic% and %lcteorologv . H rAking 401-408.Ducument 11. ,ec 111 -4. S( OS11 P IIISecretariat, National AcademN of s~lStassinopoulos E. G. and King, J. H., ''Empirical Solar ProtonS,:eIe,,\ Wa.,11I gtott D('K, 19-, pp. 119- 134. Mlodels for Orbiting Spacecraft Applications,'' IEEE 'Transactions on

"'Shea. M. A. and Smart. I). I-_. "Siniticant Solar Proton Esents, Aerospace and Electronic .Slvsemns, Vol. AES- 10, July 1974. pp. 442-19-0 19"2." lar. I 'rrestrral l'Ivsics and .kteitirology, Wourking' Doc- 450.umcent /ll. Scc. III -X. 5(1)514-1 Secretariat, Na*tional Academy of 55Stassinopoulos, L. G., -SOI.PRO: A Compuer Code to Calcu-Science,. Wkashingtoni, D)(. MN 19'9. pp. 109-112. late Probabilistic Energetic Solar Proton Fluences,'' NASA Goddard

4Arin-troiig, I -V.\. Hru-lngardt, ( -,arid Meyer. J. E., "Satellite Space Flight Center. Greenbelt, MD, NSSDC 75-11, April 1975.Obsers atin ot I ntcrplanetar\ and P'olar Cap Particle [loses from 16 Chenette, D. L. and Dietrich, W. F., "The Solar Flare Heavy-IonI9 i43to t hc lrescnt,. ti at her anod luniatic Responses to Solar Vu'ria- Environment for Single Event U~psets,"' IEEE Transactions on Nu-tions. edited h BI. M. NicCorinac, C olorado Associated lun. Press, clear Science, Vol. NS-3 1, Dec. 1984, pp. 1217-122 1.Boulder, (), I9 15. pp. I '9. " Selected Geomagnetic and Other Solar -Terrestrial Physics

' M, Donald, I . B., Solar Proton %Manual. NASA, Washington. NGDC0I, Data," Compact Disk, National Oceanic and AtmosphericDI , NASA kept. UR R-l69. Sept. 1961. Administration, National Gecophysical Data Center. B( , r, CO,

"M~odiseitle .1 . u-n _ k NI.,M. and IHardy, A. C., 'Model 1987..Solar Proton 1-~osrittents for Manned Space Missions,' NASA, "L1ee, M-."The Association of Energetic Particles and ShocksWa-,htiton. DC1, N-ASA I N D-2-46, Ap.il 1975. in the Heliosphere,'' Review5 o f Geophyvsics and Space Research, Vol,

'W\\ chcr. N\. R ., "Ani [\aluationl of tile Radiation Hazard Due to 21, March 1983, pp. 324-338.'solar P'article U1-sci- Boeing Rept. D)2-90469. Seattle, WA. 1)ec. 59L-ee, M. A., "'Acceleration of Particles at Solar Wind Shocks,''11463. The Sun and Heliosphere in Three Dimensions, edited by R. Gi.

"Wkecber, R~ .. "Suntspot Number and Solar Cosmic-Ray Pre- Marsden, Reidel, Dordrecht, Holland, 1986, pp. 305-318.dictiont it ( clc 20 165- 197;) ssith P'relimtinary Estimates for Cycle "6'Smart, D. F. and Shea, Mi. A., "PPS76-A Computerized 'Event21." Hoeing kept. 1)2-I13522-I. Seattle, WA, May, 1967. Mode' Solar Proton Forecasting Technique," Solar- Terrestrial Predi-

"Wch her, W. R.. "Spatial Variations in Solar-Cosmic-Ray Inten- tion Proceedings, edited by R. F. Donnelly, U. S. Dept. of Coin-ties.' Boeing kept. 1)2-114459-I, July 1967. merce, Nation Oceanic and Atmospheric Administration/ERL, Boul-"".Solar Prototn Nonioring Experiment, Data Set 69-053A-07C,'' der, CO. Vol. 1, 1979, pp, 406-427.

National Science Sers ices Data Center, N-ASA Goddard Space Flight 61 Smart, D. F. and Shea, M. A., "PPS-87: A New Event Oriented

Center, (ireenbelt, \il). Solar Proton Prediction Model," Advances in Space Research, Vol.k"McKinnon,. I. A.. "Sunspot Nuntbers: 16101-1986 Based on the 9, No. 10, 1989, pp. 281-284.

Sutispot Activity in the Yecars 1610)-1960,'' National Ocean and At- 62 L-ingenfelter, R. E. and Hudson, H. S., ''Solar Particle Fluxesitosplieric Administration, National Geophysical Data Center, Boul- and the Ancient Sun," Proceedings of the Conference on Ancientder, CO. L \6i-95, 1987. Sun, edited by R. 0. Pepin, J. A. Eddy, and R. D. Nierron, Perga-

"I e- rnnan. J., Armstrong, I., lDao-Gibner, L_, and Silverman, S., mon, New York, 1980, pp. 69-79."A Ness Protont I luence Model, for E> 10 MeV." Interplanetary ' v'Goswamni, J. N., McGuire, R. F., Reedy, R. C., Lal, D., and Ala,[ arti de tnironinen. edited by -1. Feynman and S. Gabriel, Jet Pro- R., "Solar Flare Proton and Alpha Particles During the Last ThreepuLstoit1 Lab., Pasadena, (A, JPE P'ub. 88-28, April 1988, pp. 58-71. Solar C'ycles," Journal of Geophyvsical Research, Vol. 93, July 1988,

'King. -1. H., ''Solar Proton FlulenCs for 1977-1983 Space Mis- pp. 7195-7205.