RD-RiZ6 29 -USING SOLAR RADIO BURST INTEGRATED FLUXES TO PREDIcr 1/1

ENERGETIC PROTON FLUX INCRERSESCU) AIR FORCE GEOPHYSICSLAB HANSCOM RFB MA N R BARRON 31 AUG 82

UNCLASSIFIED RFGL-TR-82-0253 F/G 3/2 N

a .

-U

1132~ 2.2

11111112.0

1.25 1111.4 11111__

•

.

MICROCOPY RESOLUTION TEST CHART

NATIONAL RUREAU OF STANDARDS 963-A

4

.

i ii i . .

AFGL-TR-82-0253ENVIRONMENTAL RESEARCH PAPERS, NO. 791

* e) Using Solar Radio Burst Integrated Fluxesto Predict Energetic Proton Flux Increases

WILLIAM R. BARRON

DT1CELECTE

31 August 1982 S APR 4 1983J

Approved for public release; distribution unlimited.

SPACE PHYSICS DIVISION PROJECT 4643* - AIR FORCE GEOPHYSICS LABORATORY0l-CHANSCOM AFB, MASSACHUSETTS 01731

AIR FORCE SYSTEMS COMMAND, USAF

This report has been reviewed by the ESD Public Affairs Office (PA)and is releasable to the National Technical Information Service (NTIS).

This technical report has been reviewed andis approved for publication.

0

~J7,DR. ALVA T. STAIR, Jr "

Chief Scientist

Qualified requestors may obtain additional copies from theDefense Technical Information Center. All others should applyto the National Technical Information Service.

-

UnclIas sif i edSECURITY CLASSIFICATION OF THIS PAGE ("a1,n Dot* Enerd

REPOT DCUMNTATON AGEREAD INSTRUCTIONSREPOT DCUMNTATON AGEBEFORE COMPLETING FORMI REPORT NUMBER 2 OVT ACCESSION No. 3 RECIPIENT'S CATALOG NUMER

AFGL TR -82 -0253 34- TITLE (- SobIII.) S. TYPE OF REPORT & PERIOD COVERED

USING SOLAR RAIO BURST INTrEGRATED Sinii.FnlFLUXES TO PREDICT ENERGETIC PROTON Sinii.FnlI- LUX INC RE- ASES 6. PERFORMING ORG. REPORT NUMBER

ERP No. 7917 AU THOR(st S. CONTRACT OR GRANT NUMBFROs)

' i liani 11. B~arron

9 PERFORMING ORGANIZATION NAME AND ADDRESS 10. PROGRAM ELEMENT. PROJECT. TASKZir -occ Geophivsi cs Laboratory (IJHG) AREA & WORK UNIT NUMBERS

HIan sconm Al,' B 62101F

AMassachusetts 017:31 46430:306

It CONTROLLING OFFICE NAME AND ADDRESS 12. REPORT DATE

Air For:e C eophvsi cs I ~aboratorv (uG ) 3 1 August 1982Ilans.,con A F: H 13 NUMBER OF PAGES

AMassac-htu-etts 01731 :3214 MONITORI1NG AGENCY NAME A AODRESS(iI differallf frn~o CoiIna OII-) IS. SECURITY CLASS, o 'lh, IIe1 j 000,1

Unclassified-15 CE ASSIFICATIONOANRDN

SCHEDOULE

'6 ThON PO8tTION STATEMENT 1,f rh,- Rep,If

It A~rIovedl for pulic release; distribution unlimsitedl.

C"- V 0. ION ST A TEYFN I I ho .1-v-fr onlerd mn Plor 20. if diIfe ... Iro, R,,por

IS ,IrMENIA., rNI'E,

- A- 11 , , -. -- *, 1, I ~ r -d dl,,:I,f, h, M - ,:,. r

4 It Ii !% 11f ti f * I - p o r t io n )1 )0 1Siio h u . t NI' x in c r e A s e . 'Il(,

II,-, 11111 1111!,1 IItS l th irflrtta~tt' riio lluxes :m l the proton peatkIll~~~% II Il511"nII il( jI r !lt 'ixcts wern iiultijplifrl lv the Factor

;I. 01I, ,II r I ) hI i 11L 1I It 1 it) rtIaI, 111(1115 betwkeen Ilhe site ol, lI 1:11 ' o I t I,Ij 11111 .t Illo 1 2 1 1 1 I( I i l l o ll tijoll het e' ( rI tiit Sll

EC.JNTOl LASS AICASON .D4 T.;S PAGE I '-~!l~ F,,T -e.

UnclassifiedSECURITY CLASSIFICATION OF THIS PAGE(Wh n Dots Entorod)

20. (Contd)

and the earth. The solar footpoint positions were determined from the solarwind speed.

Two-variate linear regressions were computed using the time-integratedradio fluxes at five discrete radio frequencies in the 606 to 8800 MH, fre-quency interval and peak proton fluxes at > 10 MeV and > 30 MeV. The higherfrequencies of 2695, 4995 and 8800 MHz all correlated better with the> 10-MeV protons than the lower frequencies. The > 30-MeV protons wereeven better correlated with the higher frequencies, but correlations with thelower frequencies were poorer. The Total Energy Density, ET, of the radioburst, an integration across the frequency interval of the time-integrated radiofluxes at each frequency, was found to be better cprrelated with the protonfluxes than any of the individual frequencies

Three-variate and four-variate linear reklessions using different;ssonsusig dffeentcon-figurations of the time-integrated radio fluxes'i re also computed. The best)f all the correlations was found with the four-variate regression of the threehighest frequencies and the > 30-MeV proton flux. Comparisons with otherstudies have been made, as well as suggestions for use of the results for pre-(ic.tion of PICA and polar D-region phenomena.

I

SE( R-I Ty Y C ASI Fic &TIo') qr T..rS P)AGE- rh., II.t, Ferod)

-' -- -- -"-• 45 -

Preface

The author wishes to express his thanks to P. Bakshi, J. P. Castelli, E. W.

Cliver, I). A. Guidice, S.W. Kahler. M. A. Shea, D. F. Smart, andG. L. Tarnstrom,

ior many helpful discussions and much generous assistance and guidance during the

preparation of this report. Thanks are also extended to K. Al. Flaherty for per-

forming the laborious task of typing the document, including its revision and

mn'Iifications.

- Arue S- ujr[

. 1Zj

A-5n

Dist "'

A

,,.. , ,:

Contents

I. INFIIODLC"I ]ON 7

2. -A.KG.- ':II NI) 8

3. 1) A'[A 13

3. 1 Ra'lio D~ata 13:3. 2 Proton iData 16

4. (O 1I-:I .l AT!ION S 16

41. 1 Solar' Win! Speed Corre ction 204. 2 Aultivariate Prediction Study 234. :3 'Iwo-variate Linear Regression 2441.4 Three -variate Linear Regression 254. 5 Four -variate Linear Regression 26

5. CON(:LUSIONS 28

RI H F I."R I.CNC- C.S 30

Illustrations

I . Peak Radi luix Spectra of" Two Sol ar Proton Flares, HachrShow inrg the U -Shaped Spectral (Characte risatic

2. FiT e(~tr~ enmporal Radio Flux \ariations of 281 Augyust 1466Solar Proton [tare Obs-e rv'l at the Saganiore 11311 Radlio()hsr rxatorv I1

%-5

Illustrations

3. Spectral Representation of Time-integrated Radio Fluxes ofU-Portion of 28 August 1966 Solar Radio Burst 12

nA4. Variations of the Correlation of Log (10 110) (per Van Hollebeke et al)

With Log ET for l)ifferent \alues of the Coefficient n 18

5. Peak > 10 MeV Proton Fluxes, Multiplied by 101"274A PlottedAgainst ET 19

6. Variation of the Correlation of Log ( nlIK ) With Log E for Different

V alues of the Coefficient n 21nB

7. Peak Proton Fluxes, Multiplied by Appropriate e , Plotted AgainstI I-. The straight line best fit to the points is shown. 10 \leV protons 23

8. Peak Proton Fluxes, Multiplied by Appropriate e , Plotted Against. The straight line best fit to the points is shown. - 30 MeV protons 24

9. Three-Frequencv Spectrum of Time -Integrated Radio Fluxes oftU-Portion of 28 August 1966 Solar Radio Burst 27

Tables

1. Correlation ('oe ffi cient.- 132. Riadio and Proton )ata. Radio and Solar Wind Data 14

3. Radio and - 10-MeV Proton Correlations (incorporating Van Hollebekeet at [19751 proton flux a ljustment) 19

4. Hadio and Proton Correlations (magnetic footpoints determined fromsolar wind speeds) 22

5. Radio and Proton Correlations (various combinations of time -integratedradio fluxes) 25

6. Hadio and Proton C crelations (2695-, 4995-, and 8800-Mltztime-integrated radio fluxes) 28

6

F.

I

Using Solar Radio Burst Integrated Fluxes toPredict Energetic Proton Flux Increases

'.4

The 'o,,urrence r)! solar flare activity is accompanied by an increase in the flow0 cni-r:v fronm the si', into the interplanetary medium. This increased energv flow

ieavc- the. -it in the lorm )l various emissions such as radio waves, N-rays,

1ertleti*',T irons,' jreO j m)otons.; and Alpha particles. The energetic protons, if

n (- rich ,ire .. Irth's (.rvir-onnrent, make their presence known by increased radio

a avdi.s', orprtI 1 n the polar iori).sphe re. Ihese protons also disturb some humnan

. '*t D ii s; ! x' T.X l in peedhin fr' riolspo r I r1 dio comnll-n icritiOns [called polir

,p i-,r p~in 0I( I ll, r,,iriptre iotiispheric D)-region height wnich upsets some

li) ti % '. ,itior I lrt ti 'f Il, in (tl Uis rupti bot ii me on.d a rid inniann2d space ventures.

r, - .I I i r tit Ir + sr l ,!. I ,tidi (-d (Ilistrbirbices aire (f sufficient ivirgnit ude

-is ri e' -. 1- r 1 i) ),.ever 1lhstrilti-in ,smirrs ind 'auise problems in teler)hone ,it,-

1!11t

n tt I Irrl isrl-4 i-Jili -Lli ti'ihtl'iIr S trs. S tich as the lrans-Alaska

i'Ir . "1 sie dii I eli dlStl hari' es to inan's activities have pr)vided an incen-

r ti" iu1v till I -; zi . ti irv, ziwl the inipa .t of this activity om the terrestrial

; l : ' '* - 1 : ' ' it I ' l l I )lt .. A , IIL II- l I l Q };

I H, i\ .11. 1 t ;i I , So 11ulr- Ierr'srri:il ' ie tions liroceetinlrs, \ii. 11,R'e I, 'ts in , RII'v It( h rI Ir ati it Itdar-l c r I I lri:I -

ii "-t:rir ]j' '~T'el1T iJ;HT--g -7riL hirll r, ( o rio t(lo. 23-27 Airril 1'17!1.

... 9v . 4fl - - -*- --- , r-- "-' i - "- - - .rr ,.I r .- i r , r*. 7 ' ' :-- " ' " . " ' " " '

The purpose of this study has been to find potential relationsiips between tie

maximum energetic proton flux, attributable to a solar flare, and tie radio emis-

sions from that flare. Tne effort has been directed to finding adjustments to the

peak particle flux, as observed by earth satellites, for the position of the flare on

. - tile sun, and combinations of tne radio frequencies observed, in order to obtain

better possible relationships.

r2. BACKGROUND

In relating solar radio burst phenomena to energetic solar particle emissions,

5 Castelli et at and Castelli showed that a particular radio burst peak flux spectrum

could be associated with the occurrence of PCA's of 2 dB or larger. The PCA's

are related to the increase in the flux of MeV energetic protons. The radio

spectrum was found to have a U-shape when the peak radio fluxes in the meter through

centimeter portions of the radio spectrum were plotted. The time constraint for

t this condition was that all peaks must occur within a few minutes of each other.

The minimum of the U-shaped peak flux spectrum occurs in the decimeter portion

of the radio burst spectrum with the peak fluxes rising to > 1000 sfu (1 sfu = I solar

flux unit = 10 -2 W m -2 Liz-1 ) in the meter and centimeter portions of the radio

spectrum. Figure 1 illustrates two U-shaped solar radio burst spectra associated

with energetic proton increases observed in the vicinity of the earth.

Several different approaches have been used to find possible ways to relate solar

radio burst parameters and energetic particle fluxes. Straka' studied radio burst

parameters such as burst peak flux, burst mean flux, burst duration art burst-time

integrated flux or energy at five frequencies in the 600- to 8800-MHz frequency

interval and related them to the peak proton flux of the associated particle event.

He found that the time-irntegyrated radio fluxes gave the best 'or'elations with the

proton fluxes. lie also found that the best (-orretations were aith tire lower fIre-

quen,.ies in the observation interval even though his sampulei szes riiav h ive been I()o

strall. In hlis integration )f the radio fluxes, Straka intezr:1(''I the .rtir. ;iux

increase ()f the burst, indludingr any associated post htrrst ir riase . N 10 'It

ilit*ig rt tin \ st e ,, -ti-Iixirirni p l0ion ()of the di i f111 II t V. ri " J :" -,,. 1W

tir,' ra1di) Itir it rinies. Ie oilnd th;r tile best 'lirrel.itliiris fwfi#. I Iaricl e t(I

rmtlii ftirx(es ind peAk prriiton flixes wire ;it t c iigwr trlr,(114 t<I( (it L i trc i .it,

h;(nd sfildled. ie is() toiund fil;nf ;dit s lt it" Lie' bsi ' ,fl ,, lix 1i- '1-.

po"is ion ()i " rrr 1sit ti S )ri flZre )t fti ff i. ,i i tt I it nt e iS dr ,. il.l 'i,.r( 1.r fi (. e r', . rr1ir11).r. ()t Ci td ih i\ . ', hi I dSe s )1(

('liver related integrtied burst fluxes at 2. 8 Gliz to PCA magnitudes; fe did not

find tne flare position correction to provide any improvement in tue correlation.11

It is interesting to note tiat tue 2. 8-Gliz frequency used in the Cliver study falls

in the decimeter portion of tne radio spectrum; this is the minimum of the U-shaped

burst spectral shape found by Castelli et al 2 to be predictive of a proton event

having occurred. Akin'yan et al 12 studied both rising portion and the whole burst

integrated radio fluxes at frequencies of 3 and 9 G~lz. In their study they also

made adjustments of tie proton fluxes for the heliocentric position of the flare.

100,000

29 SEPTEMBER 19682B FLARE N17 W51

q

1000 / 1

100 . ..... Figure 1. Peak Radio l,'luxtoo 1000 10,000 Spectra of Two Solar Proton

S00,000 lares, Each Showing the- U-shaped Spectral (Characteristic

Z23 JULY 1970zz W 1B FLARE N09 E09w

0

x

- I 0,000taI.

1 00

100 1000 10,000

FREQUENCY (MHz

I1. live,'. I. I. (1!76) Parent-llar' IK!:i-si, m at '2.8;i (;l as a ll',',iit) f th,Peak Ahsorption of Polar ('ap Ivents. NIRIIl H -IS. N2 \ ht ,I.abratorv Center, San Di(,ro, T'71T7,)ni, 1'2152.

12. Akin',van, S. '. . Ion ich(, . . , ',. a, . I. J!". (1 '7 ) I'st i' , 4 '

itensitv F .s4olair prt n olue, oli t r)'( ' ;L 11 1,)! W 1 ' '

hursts. (W'W )rfla r lt . Aml'l. 8 rNo. -1):

-.. .. . . . . .. .* . .. r - y > . '-> - . .

K. All of the studies mentioned here were made at discrete frequencies in the

decimeter and centimeter portions of the radio spectrum. Croom13 also considered

several different time and flux characteristics of solar radio bursts such as peak

radio flux, mean radio flux, time duration of the burst, and mean duration of the

burst as predictors of the peak proton fluxes. He found the mean duration of the

burst to be the best predictor. The mean duration of the burst is the time-integrated

flux or energy of the burst normalized to its peak flux. One other characteristic of

the radio burst to which Croom 13 referred was the total energy density, ET. A

representation of E T may be found by means of a double integration. First, a time-

integration of the radio flux increases is performed at each frequency of observation;

trin the time-integrated fluxes at the frequencies of observation are integrated

across the observed frequeniy interval to give the total energy density, ET. of the

radio burst

Bakshi and Barron studied the use of 14ET as a predictor of > 10 MeV proton

flux maxima using integrated fluxes at frequencies of 606, 1415, 2695, 4995 and

8800 MlIz. Instead of time-integrating the entire flux increase at each frequency,

however, they time-integrated only the portion of the radio burst flux increase that

contributed to the U-shape spectrum discussed earlier. In some instances this

time-integration was of the entire burst flux increase; in others, it was not. Fig-

Ures 2 and 3 illustrate the procedure used. Figure 2, an illustration of the solar

radio burst observed at the Sagamore [lill Radio Observatory on 28 August 1966,

shows an example where only the portion of the burst exhibiting the U-shaped pro-

file in the maxinmim power spectrum was used to obtain the time -integrated radio

fluxes, First. the time-integration of the flux increase at each of the five indicated

frequenoies was found. At each frequency the trace of the flux increase was broken

into a ( ollection (with a small amount of smoothing) of plane triangles, rect angles,

and tra~pzoids (Figure 2). The areas of these plane figures were then found and

added toizether to give the representation of the time-integrated flux for that fre-

• uencv. I lie five tire-integrated fluxes were then represented, as in -i pure 3, to

pive [our, trapezoidal figures. The areas of trapezoids enclosed hv the adjacent

trequencies were fourld, ;irc1 finally tile four areas were added totgether to give tie

I.:, for the radio burst lsing ionsidered. The use of electroric digitizing uiro.il s

and corllputers shcold 1rrk it possible for this ptocess to be more a(c-uratclv and

r:cpicltv icerfr ned ihr th, 'tlture. l'ht'Se -is well as tie tine -intearatel ra(dio

flux at the live frrqrirntics ,%ere I herr correlater aith tire peak hourly

,,vrage it" h0 \ iroprol flrrx ssccciati'd ith the rwit 1hurst bring >c,! sitrel.

3Id. (roonil, I). I+. (197.. ) rrrcstira the irterlsitv of solar proton c'vents f rri thretur i rarctrsis foc)I mlal. i.rowa.ve brrsts, Solar Ilvs. 11(No. it:17 .

14. llakshi, 1). . anl !airun, It. (1 78)t Prediction rf till Protcr -ilx Iarrniturdusfrom Radior lirst rI)ta, I 1- - - 00, :\i .,057-7 -.

I1

| l

3 30

4Y95 MHZ

3.,.

2695 MHZ

1415 MHZ

2

60t, MHZ _90

I iv Ivnoa H doFu ratos f21Au it16 oa

I I, f*1()s r o 1 eS La o c Ili1 a j h v ao v vtep aI~~~~~~~~ ~ ~ ~ ~ ~ tx. ,!iI i n --ihtdat sae pc ,11

28 AUGUST 1966

ET -LOWERE-A

.~45

o40[

0

Figure 3. Spectral RepresentationIt of 'rime -Integrated Radio luxes

0 30 of C-Portion of 20 August 1966< 5 , PE Solar Radio Burst

'5,

10' 1

- 606 1415 2695 4995 8800

F REQUE NCY(MHZ)

Table I gi ves some r ' the correlations found. It is seen that, in the case of the

lot ept at ion )fI mu ivida (01Fr-eq uenci es, the highest Frequencies gave the bette r correla -

t onrs. With the c(c .ptiol 'if' 4iWh S-M[,t7 data, the correlations of the indlividual time -

in tep rt'aI radVio I ;e1nd F"T were futer enhanced hy the adj ustme nt of the oh -

se r'velopak hourclv ave taged pa rtiec e fluxes fo r the location of the sou rce flare on

he la oIl he un. 'I i' (11 ustine nt factor usedl was e' , where A is the tongit udi natdi sto (14, in rai~i ans, roc tFie locati on ot the flare to the solar tootpoint of the i-np -

netii' till lines, conne(tiW, tIre- earth to the sun at -57' WV solar longitude.

FBrrrarr rld Bakcsh Fr ifound tltat a bette r correlation hetween FT and the2A I3A 1

IIt) -Ai\t ;w l rri{tat i -Ic flux asobtained by trsinop e rinsteadl of e . It has also

Fo'rr toundr I that het( o id lustime rt 'tirthe r im proves the correlations at the top

Fr rec il jivi. niil tr'equnrries it] the fr'equenc-y interval considered, as showxn in the

4 ,t ,a hw ri )! Fabrle I.

IT will la' tire tpurpose' olt is report to discuss the effect of adtustitng the

lO),tIH)t it po i tir for diffirerrt soliar wind speeds, arnd rdifferent combi notions arrl or

'~i'iprt ()I 111the tinre-ilrtepr;rteF radio tfluxes.

5. 1 ,t I.rI.,tjr, .H. , 11t1liFirkshi. F'. (1 "7 !)) A tpl i o nrr ni Irtegratel lHidio 1311rstI~ristor ire( Predictiont of, Srl1rr' Filrt('wlei( Pr-ototn Flux lrc'~ssSoar'-

FIr Ptra1redictionrs P'r'oeerlinps, \ Il. Ill: SrAr'.tivitvltedtorsltttn'tI.'IItr Itinl oa''e t' estHi ,iT Idicr l- -''e -,, r- ;rd\tc~lrr

2:1- 7 Ap il VT 1 i

1i ablae 1. co~rrelation Coefficients

Proton Variables

Observed Adjusted for Position

Radio 3A 2AVariable-s 10o 10 e10P

A

Timeo-integrated flux at:Ofv liz)

8800 0.718 0.758 0. 779

49915 0.781 0.776 0. 820

2 69 5 0. 759 0.777 0. 802

14115 0. 601 0. 749 0. 734

606 0. 53 6 0.707 0. 682

'4606-8800 Nllz 0.753 0. 806 0. 824

31. D)ATA

The sola r radio and energetic particle data used in this Study is given in

Tale 2. i'hese dlata were obtained from nume ruus sources suchI as Solar- (Ueo-

physical Da1ta (SGD) Reports, Quarterly Bulletins on Solar Activity of the Inter-

national A \strionomical Union, and vaIrious ubse rvafory reporets.

:1. 1 R~adio Dataa

Thec raio dlata are primnarity dlata taken at fr'equencies of 606, 1415, 26q5,

-1995, an(l 8800) Mllz at the Sagamore Hill Radio Observatory, the Mlanila Observa-

tory in the Phili ppinres, and the Athens Observatory in C reece. These observations

providedI a eons iste ot long-timeC set of (lata at the f'ive frequencies m entioned. The

4 e~pak hi x radio hun. istpect rum of all of' the events (cons idered had the U -shaped

,5pec1 u df01ls err bed by Casteti et al. 2 In two of the events. 1 and 20, radlio data

inf ii9(m t00 rom iberaoisother than the three mentioned werie iized. In

f ,ent itO m 1, 2-4 Akl reh 19 66, no observations; were male at the !Iye i ndicrated

freqierlik at the tinme the flare occurred, -0225 UT. In their p1lace the re -

portedl vAlues at 500) Al lz f'rom iiraiso Observatory andi 1000-. 2000-, 3750-,

and110 )-O-Aliivzie from 'Nagyova Observatory iii Jlaan were uised. l,:e(,t 20

a s iwfs b% bhe AtloensOhb,-rtorv %% Ih1 'lot, lntobse rve at ie 606-,NIlz requenev.

-dL N o toM0 0 w 0 to t cr.

- cc 0 c C rJa C; C; -C; 00 Lr 00 C co Cq C;q- M 'l-- C'3 t- 'co 0)co

- r-

00om

C,3 m 0 oNmNmww- ,3 o0mw

Cd C D - c ';(zLr . f;06N 0 C' r ;f 06 r: 14C o-; l

0 -- o cNC) I t o 1

0 0c

Ig A Cd

a) Ce o C)ODtot

Cd U

m 0J Nl o 1

0 MO -o tO 0tO.C 'Je mJc )CD M Cq en-n-co

0 d 0o co co N

m C, v mn c'] n-rmto - - c m m

C..CD

CU ~ ~ ~ ~ C: N w . c o cc- 0c

Co) M-- 'a- )- Coac - D0l 1

CU CoI m 1 I - - I -1 CN 0

m t r- a) ,cm i ,Co M( t oC o t I -I -r

tX O >

r- mU wA Dr ,mI D rr "CiN.0 CU 1 1 1 11 10 10 c ci n

-U r. tn cT, ~C

H CU .2~ZZZZ4

Table Rad~io and Proton D)ata, Hadio anid Solar Wind Data (Contd)

Radio arid Solar Wind Data

Inteerationflof* T ( )sAceross the d VIn~dex 606-8800-AI-Iz Frequencyv Internal(' solar wind 404/ solar wind

1101.6 458 (1) 0.88110. 6 4431 (1) 0.91

q;17'5.' 414 (1) 0. 981 17 Vi. 0 3 64 (1) i. 11

a12 1.8 640 (1) 0. 6331. 5 :37 2 (1) 1.0!9

2.2570 ()0.7 124. 1 470 (1) 0.865Y 7 687 (1) 0. 59

.n1154 1D 0. 891 II I457 1) 0. 88

51 1 .157 (1 0. 88* ~150 (1 0. '40

18 1 0. B83lii 10 41 (1 0. 83

N i12.14.0 804- (2) ). 50t2 ;2. 5 61 (21 0. 72

18 Ii. ii6631 ( 1 0. 61ii 3 85 (3) 1. 05

2 1,2. 66G 201 3) 1. 26

-11 (1

r f I.i n , MTi 1 )

III it- I( )1-u thi h~ itorm Itiol at '536 Mliz ob)5lved it ( )rtidlejov, (ieecioslovakia

'A as ios. A he Irti-L L(t' raoit Iluxes for all the Ireqlten( i(5 o01 rvent I mid the

-36--AilI'fe'pwtit v 0 1v 20) Ve(e calctmot~d [010 thle evetit rlafi F'lux miultiplied

yh ! li en'rt dI LIt I'M 1)1 1 ,- Lhe aolo'' I ra- os W' theS- (eents were riot av l able

'o tIt10at1oi' om OwI vp( s-i ' ,tthe earlier. 'I he( last three, flares mo the aible, ill

u-nre alter I, 1 127 eutTor b l iii~~ etween solit e-vcles 20 aold 2 1. '1 %o oft I hosu

20 arm, 1 2o 21. kerco bexlo , cur-n at 240- N so)lar l atitIude, whi, 11 'A ld

,i ', th''t ninwofm nlh h-a 'I yule1( 2 1. ;<lare 1I w s )e red M a 07 N ola r

ititut'Ie. rrfrit Ins tokihlv siga %([ilt 08 Ie 20th SLIIr5 1 )Ot -v('.

3.2 Proton Data

C The particle data consists primarily of the peak hourly average proton fluxes

at 10 INleV and > 30 1leV as viewed by various Explorer satellites. The first

- three events and the last three are reconstructed from other energy channels, as

10 Ile\ and > 30 MeV channels were not available when these events occurred.' 'The particle flux values used were adjusted bv subtraction of background particle

flux levels and, when the increase under consideration occurred on top of a prior

increase, any flux due to the prior increase was estimated and also removed. One

additional possible limitation on the observed proton flux values was pointed out. 13

1y Coorl where he indicated that the true particle flux peak may not be well

represented bv the hourlv average values being reported. When the averaging

process is considered, it will be realized that the only instance where the average

,fa gr'oup of values will equal the peak or highest value in the group is when allF alues in the group are equal to that peak or highest value.

t. CORREILATIONS

Observations :f solar flare phenomena by such means as multifrequency radio

hurst tneasrrrements, Il-alpha and optical observations, and X-ray measurements

all provide information about what is happening in the vicinity of the location of the

tlare. The radio flux increase measurements give an indication of protons and

electrons that have been accelerated at the site of the flare. If this location is a

,distance from the footpoint of' the oragnetic field lines connecting the sun and the

carth, there is goig to he a decrease in the number of flare accelerated particles

that will reach t tic sorn -earth niagnetic footpoint. There will, therefore, be fewer

1 r>articles to escaple into the interplanetary medium along the field lines to the vicinity

f the tEar'th. A ttenrpts have been ttade to compensate for this particle loss in

riorr-; prelictitnr studies, as mnentioned earlier.

\ an lholheheke et t 1 notedh that the peak 40-A1e\ particle flux observed bv

e.arth sattillite's appeared to decrease by two or'ders Of r tazrliturle for' every radian

4 e,,twen tire Ilare , site atnd what they referred to as tire 'preferred-,onneetior'

rftZiot. Ihis (trresiponds to tie p''O titlrurl Intitutrlirral interval'' (O)D of

Akir'varr eft I but is not necessarrilv tire ''fast propagation retrorn" of Heinhard

16. 1 tn IllI-heke, N1. A. I., Ila Sttn ,. 1_S., aid t \ l)'ri)na1l, F. I. (1975) The varia-lirn t s olar ir(tor erer v spectra and size distrihutior with helioloritituhe,

Srvrr lhv.t. iNr t:It'.

I. Akir virt, S. I., l"rtnichIe'v, . i., attd ('lirto k, I. 1. (1977) )etIrninaii tort ttIre ratlnnters ''f solar prrotonts inr thre nreighb)rhnod oft" the earth frani r';diht.r 1s: I. Irtt,'tsitV Irr1tltt'ionl, rncrjj'Qltr't. Ae\ rott. 17(C r. 1):,).

I I

-18

and Wibberenz. 18 With the use of data from Table 2, a comparison was made with

the Van Hollebeke et al 16 02A coronal gradient attenuation factorto see if some other

exponent coefficient would provide a better correlation between the observed

> 10-MeV particle flux and the double-integrated radio flux. The adjustment mul-

tiplier used with the peak > 10-MeV proton flux was 10 nA , where 0' n ! 2.5 and A

was the solar longitudinal angular distance, in radians, from the location of the

flare to the footpoint of the magnetic connection between the sun and the Earth,

taken to be at 570 W solar longitude. This is approximately the center of the

"preferred-connection" region located in the 200 to 800 W solar longitudinal interval.

With A determined for each event, the problem was to find what value of the

( oefficient n would give the highest correlation between the adjusted > 10-MeV proton

peak flux and frequency-time-integrated radio flux, ET.

The variates correlated were:

X = Log E, the independent variable, and

Y = Log (10 nAI ), the dependent variable.

Ii0 is the peak hourly average > 10-MeV proton flux.

l.'igure 4 shows the variation of the correlation of Y with X for different values of

the coefficient yr. The highest value of r, the correlation coefficient, is seen to

fall in the interval 1.0 < n , 1.5. Solution for the maximum of the best fit parabola

to the six (n,r) pairs gives n - 1. 274 and r = 0. 850. With use of this value of n, the

ad.usted I10 fluxes were calculated and the coeff.cients of the best fit straight line

,'iteriinel to) give

'I' = 1.815 X - 1. 17

whe'e X .ani ' I as liven he re. 'This nrav also be written as

0 2 - I. 2, \ . 45l0 ( . .3 ) l0 1 ,.

'.,he ' I " . -Iat lie (s 'veHilo 10-Me\ protot Flux. Figure 5 illustrates the

)()int rrii I ahle, 2; the straight lire is given bv the expression above.

The titnie-iritegrated radio fluxes at each of the five frequencies being considered

IIi this .tilv nI, i- o r erelated with the , 10-Me\ proton lux adjusted as nten-

lnine . le r(s tilts W)F thres' ,e or, elatiets andu the coefficients of the hest-fit straitt-

limw deter ir, tri , tile -lta et inI Table 2 are given iu Table 3, along with the

18. Ib(.inrhrar , N. , rtl I ilubr-iriz, (;. (17.1) 'ropag2ation oif flare protons in the

slar atni-jile. Solar hvs. 36(No. 2):473.

17

: . L 7 : i - 7 L - : } . . .' S: F "

aforementioned E, correlations. It is noted that the correlations are higher at

the hip.her three frequencies than at the lower two. In addition, the combination

)i the time-integrated radio fluxes. 1 ,, 1 correlates better with the adjusted proton

fluxes than do any of the individual frequencies.

FROM BEST FiT PARABOLA0 85 -

r0

084 '-

O8o

3 1- .

2 73}

O 7.'

0 0 5 0 oIS' 20o 2 5 30

,,[ l~)a (101 ,"\ I 10 ) ( pe r \ an Il th 1 k

t, 1 ' I II I o ' , I :, I )1 I I ti ; I ,( ti l i cI I I T ', itl I i, litTI I . I ill ' t I ,l -! , II - , ' pr, ) n I ILI ; .A I. tI ,.

) it 1, , -iT iT )1iT ,1111 h . I A ho I)c jitl "

.'iO Mev PROTONSI r

p Figure 5. Peak > 10-MeV

Proton Fluxes, Multiplied

1. 274A, -lotted Against2 T . T he straight line best

fit to the points is shown

00I °

10 00 1000

E, 606 58,(.% MH,-

Table :3. Radio anid > 10-. IeV Proton Corre L tions inco'poratzngVan Ifollebeke et al 19751 proton lux ad.lustwent)

Frequenrv Correlation Straight-line (Y - nI n b) (oeffitients( kllz) C oefficient r. b(i nt r -C (t ) 11(slop ,e

6800 0.804 0.375 1. 677

49 5 0. 806 0. 4.| 1 1. 00

265 0, 803 1. 035 1. 875

I115 0.753 1.782 1. :55

606 0. 703 I. ;06 . !08

606-8800 0. .58 -1 . 1'? 1.45

1!.

4I Solar WUil Speed Correc ion

The discuss ion thus far has as sumed that the tootptoinlt of' tihe iiiageti c fichI

lines connecting tile sunl and the Earth was fixed at approximately 57 degrees

(1 radian) west of thle solar central mreridian. In reality, the tootpoint of' thle con -

niecting field lines will v arv in position relative to the, 4()1at- central mneridian with19

changes iii the speed of' the solar wind. Fhe higher the velocity of the solar wind],

thle closer to thle solar, central meridian will be the footpoint of the miagneti c field

lines which ace frozen into thre solar wind( plasnia. TIhe relationship between t ht

solar- wind speed V Skin kilometers per' Second observed at thre Earth and thre

long1itu~dinal posit ion. o A in radians, west of' thre solar, centtral meridian as given20.'

1w Sm a rt anti Shea is:

4040A \

SW%

T1his e xpress ion was used to find thle diftferent f'oot polit io~si ti ons tor f it(- eve ntsa

listed in Tlable 2. The solar wind speeds used were thle closest speed., precedinrg

the tinme of the solar flar'e involved, as given by IKing. 2 1 Th , lr ala'wn

speed was used because present cons i die rat ions a re for the pu rpose of predti cting

hie proton flux increase front a solar flarie. It' thre prediction were hel 1u 0 for the

changed V % due to th- flare, thle enrieigeti c particles mnay atlready have ari'ved,

inaking thle pr'edliction unnecessarv. 2

Trhe adjustini t factor used in this case is e n , where nt is a constant to he

dete rmined and 13 is the angular distance, in radia n,;, between the longitude of thle

flace and longitude of' thle footpoint fotund as described ear'lier-. The procedure was

tie samne as in thle previous section: to find the value of n wvhich pr'oduced the host

correlatio i, and then thle coetficients of the best fit straight linte to the data set,

Vattics of' n were obtained fromt correlations of' thle totil enerigy densitY, ET.1 in the

606 to 8800-Allz fr'equencY interval with adjusted piroton fluxes at both I10 -MNe\

and ,30 -2e% eniergy Levels. Figures 6a and 6b s iow thle cor'reltat ion coeffIicients

~)btainedl for different values of n for the > 10 -Ade' anld 30 -\l\ protons. -c -

spvctive lv. The values of nt t'r maximnum correlation from thre two plots are:

n 10 2. 682 tot' > 10 -M.e% pr'otons.

130 3. 294 for > 30 -Me% pr'otons.

M ue to the roto er oif Refe9'r'ences cited above, I eyttill riot be liste fi re eSee lReft er-e a, page :3 1.

2)o

0

z00 FP

0 -

08B5 r

0 .80~ FROM BEST FIT PARABOLA

o 0.75

Figure 6. Variation of theG 70 I 1 _ Correlation of Log (enIt) WithS0 0. 5 1.0 1.5 2.0 2.5 3.0 3.5K

0 0Log ET for Different Values of the

(a) Coefficient n. B is the distance,in radians, from the flare site tothe solar footpoint of the magneticEarth-Sun connection as determined

085 -'30 Mv PROTONS• from the solar wind velocity.Figure 6a: K = 10, that is,> 10-MeV protonsFigure 6b: K = 30, that is

FROM BEST FIT PARABOLA > 30-MeV protons. The points ofmaximum correlation are indicated

00

' 075

070

650 05 10 15 20 25 30 35 40 45

(b)

By i ncol' Idtion of these n-values in the e adjustment factor, straight-line fits

were made to the adjusted proton fluxes and the T 's of the burst and the five time-

integrated radio flaes. Table 4 gives the results of these straight-line fits. The

three higher frequencies are again seen to correlate be'er than the two lower fre-

quencies. It is Also interesting to note that the > 30 MeV correlations are slightly

better att the top three frequencies than the - 10 MeV correlations. Plots of the

best fit straight lines and the points used to determine them are given in Figure -

and ligure 8 for - 10 Ale% protons and > 30 MeV protons, respectively. The

equations of the two str iiglht lines are

21

27.

> 10 MeV > 30 MeV

Y = 1.696 X - 0.991, where Y 1.642X - 1.413, where

X = Log ET (606-8800 MHz) and X = Log E T (606-8800 MHz) and

Y = Log 1102 682 B Y = Log I 30e 3 . 294 B

where B is the solar longitudinal angular distance, in radians, from the location of

the flare involved to the magnetic footpoint location as determined from 404/V S.

An observation on the results obtained in this consideration is thai the valueI 20.n = 3 used by Smart and Shea, is between the values of n found in this investia-

tion for the > 10MeV proton fluxes and > 30 MeV proton fluxes; for this reason

n = 3 can be used with little degradation of results for the correction factor in botn

crie-rgy chinanels. It is also a simpler number to use in any computation scheme.

Tatble 4. Radio ind Proton Correlations (magnetic footpoints determinedfrom solhir wind speeds)

1: requencv Correlation Straight-line (Y = mX + b) Coefficients

(Al z) Coefficient r b(intercept) m(slope)

> 10 AleV protons

1800 0.802 0.490 1.516

•905 0. 815 0. 529 1. 742

26')5 0. 780 1. 125 1.64)

1415 0. 761 1.6813 1. 266

606 0. 703 1. 781 0.823

606-1800 0. 86 1 -0. 991 1. 696

10 1i('\ protons

:00 0. 1;25 -0. 064 1. 529

".)5 (. 8;26 0.007 1. 730

2W'5 0. ":::1 0. 610 1. 625

1415 1). 1I ; 1.2110 1. 140

6 606 0. 646 1. 320 0. 742

606- 8;80(1 0. ,5 0 - 1. 4 1I 1. 642

22.

S

0C6

10 MeV PROTONS

5100

_ ",

IOM

00

• '00 0 e0 10000

Figur 7.Pa rooOlxe.Mlipidb

b J

L gO0

0>

-O *00 I0 O0

atd ha te es rlaisigur beteen irotegratues, Mudipluxsai jsed yLf(ti prtonMixspwreobtited e n Pthed totainerg enst TETinteosrd

straighttr(Iie~ bane best used.h pitsI son

• snk2 Mtiaat inrdicted tha ntkndtefvyrqenisivledidvdat

Thh , kp treelts btthe correlations erfre inhe pto flues.edi section, ni

'.ar'ion-4. n,u~tivariate combinations r)t" the integrated fluxes at these frequencies will

he( mlade, to provide linear fits to the proton data.

r23

II11

1 10 '00 1000 10,000

E, (606-8800 MHz)

Figure 8. Peak Proton Fluxes, Multiplied by

Appropriate e13, Plotted ALainst lIT* Thestraight line best fit to the points is shown.> 30 Ale\ protons

4.3 1wo-%arie linear Regression

Two-variate linear regression icr the iam iliar straight-lire fit for finding the

coefficients of the expression Y = m.\ b. This two-variate linear regression was

tused in the previous sections emplovioTg onlv time -intearated radio Ifluxes at dis -

crpte frequencies or the total energv densitv . , as the indiepondent variable and

the corrected peak proton flux as the dependent variable,. Here we will investigate

some different combinations i)f the time-int egrated radio fluxes to provide varims

independent variables.

In reference to Figure 3, lthere --ire four trapotzoidal ar('as indicated, two oi

which involve the lower three frequencies, and two of alich involve tlie upper three,

frequerncies of the five use,] itt this studv. The Ireeietwv 2695 \Iz7 is secn to htt

irIluVeul in b)th the upper and lower frequtlericv .ets of trapezoids,. I ronii "iL, 'r" ,

| 2 4

I

the areas of' trapezoids 1 and 2 were combined to give an E T-lower and the areas)f trapezoids 3ad4 were combined to giean E -pe.The E Tlower and

1T- upper for atl the events in Table 2 were then used as the independent variables

to o)btain the coefficients of the best fit linear regressed expression. Again the fits

were made to the logarithms of both the independent and dependent variables. S3ome

its also were made using the actual values of the data. The results were not very

different as fat, as the correlations between the variates were concerned, but the

oefficients obtained were much larger numbers which would be more cumbersome

iio utilize in a computing scheme. Table 5(a) gives the results of the straight line

i t t EF loerandH,-uperwith the)eak adjusted proton fluxes at>10 MeV

indI > 30 IMc\ energies.

Tlable 5. Radio and Proton Correlations (various combinationsofI tinw- -integrated radio fluxes)

(a) Two-Variate Linear Regression. Y = mX + b

Low I- rvquency High Frequency

> 10 MeV > 30 NleV > 10 MeV > 30 MeV

0. 807 0.752 0. 826 0.839

1- 0. 97 1 0.571 -0.874 - 1. 397

W1.3 22 1.208 1.7 13 1.707

(b) 'I lree -% ariate Linear Regression. Y 0o a1 1 ,fa 2 X 2

I ow F rquenmv High Frequency

10 .I(,\ 3io Nlek 10 M\e% 30 IMe V

0.12(10.792 0. 827 0. 842

IL11 0.544 -0.446 -1.065

~l1 -flic-0. 187 0.4-47 0.1(30

I. '1i50 1. 74 9 1. 2 6B 1. 530

Inistead if ,omhtirrninL thti a I areas, 1 arid 2 w I,1 Tlower and 3 arnd 4 f'or

I , inl Figuri .1, 'k l10?-2id i neilar 1i'L'rssior) lit (0,Ole for-

4~~ ~ I~ X '1 " ) A" \IL., 1i1i' rn .l ariates in the cxpression at,0 1 1

2 5

0T

* I l). CI rapelzoid: I area L, uppe

WE oni (I rapezod 2 area Qm01 H' &we r)

I 'ii ('I rapoiid 4 arca fron [:, F -u piner)

t- il I i I, ft i aA- I ut IAdi e 2 ; oii irO j(I( t hAe hi- s redl ,I t I [at e's, we ob i ill lt'

5 IAA~h'Aff- 2.nef in 1[1lic 5(1h) (Ansiderati,rs ol* I :ibie 5(a) andi 5(h) pi1la

lhei thri ' -e-varit ( f muresii OVes a bietter Htl to W e dmiA> set Man11 Wi 1 wo am

t'Js-iti. IiThe 11hle It~jef)A'OCES d~ also h''btter 'AL01 th hA' ~ Iw i r'i')nrl Vs,

*iS'1' l+t joied h%, the correlation roeri'Nipints. to the > 30 1\lnV pruton data than to

11A' l) M',l '. oor p110 ata.

t.5 IAAoI-%mm .AIt sA' i iAn'r Rvit' sliAA

.ii[t .l- slilo- 1 I nlf pr'Otlol (Iota, \tA'li bif~koli apari 1ra it ''A !our- t'

[A f11' IlI i' I'A if f', - I Al A th JIA III1

~~n-I r - i f t 'Ito nl0 lx)

o Pl '7) 1 .~ r c'ti vie, . !Lur I

I i it-,- i. A l if- A tA o*(,' 1- : PA' !"),il I hv hIV ''Aill, j'I . P :

:t I'l, A11 l iti l' l '1,11 r pl( l'',,l' p l fill \k~ )I It',, ill t i t ~ ~ N l

;0 i (l l!" %''ii' )-j~A '. t' ( I A the rec1tarvtIA- wer. A1,, c- ' '

i UPPER-MEAN FREQUENCYIKILOWER-MEAN FREQUENCY 2

.20.0 F UPPER-MEAN FREQUENCY 2

NF I LOWER-MEAN FREQUENCY 3

U. -UPPER-MEANo 15.0 FREQUENCY 3

w. RQEC IZ 3

FREQUENCYC (MI z

J.igue The-irq c Sp fIieInerae Radi Itxe ofUp nof 8Ags

'I he(1:11rew 1 guwe th. 1 herea frequency Sp andru he of im a frqec

'Ih( u~r r,-(I(-iused i ) our-vritinearh rper-essin areqec ifrltelicatedli

i(, w is 15,40 M~tz k~l f~ is the next highest frequency obs erved at the Sagan iore

II ill Radii)obecvtic and the oither hiseocvatocries nf the Air Weather Se cvi ce s

Radio S'ilar 'l'csoi Nitwocrk (IIS1N)1. 'Ihe best fit linear re'gressio~n oefijeotts

wic.th'i 'ibtIcIItl LVsing Owc data in Fable 2. Table 6 gives The re-sults of this cc(-

ciso.'1 1' oreilionl ''oef~i ijent, R{ '.)24' is (allel the tiultit~k,lciiccelatiin

m. i Tit 0I o~rIii I .i~t'.'-w 2, 3, arid' 4. ' tese( occes-poni to Y, \Itiit . InI tie f-i wi r'<sinr .'I th, Iincau lit Y~ - ao A *I. , I :I %, Iit

1. ' 1 1 1 2 3

S ii 2: %%27w ild hi .. xrw 't I)( itt 'ircilatinir -wftlieit hik is- 'i iso,'.' it.

IIIm Al. .''I' I

V aitt G. Radijo anta P'roton: ( tcelatioi's (2695-,.19 t-,z II I 880tt -AlI I t i I n -iItegtatcd al ( t iti flurxesi

I'our-\ artate Regr'essionut: - : 0 a 1I, 2 a3:1

> 10 NMeV 3) 0 Mec

LNI1 2:34 0. 87) 0. 895

S 240.522 0. .7275.10-0.78 1 2

a-0. 3 94 -0.684

2 1. 293 L. 335

.,0. B0B 1.023

'4 Smile ( osidtoattorts have heen tmade here or UhV relationshi ps between the

ic irated raioi fluxes of' solar flares and the associatedl energetic proton flux

IrICtase0S. Adjustmoents to the lmt-ttiIck flux for the Imon ~ofl Wte Hare v"t the tue

I il( situ1 hare heen applied and various iomihinations and weightings of the jute -

r.:Itcl radio luxes tare alsot been tried, In the end, if' to correlation coetho,:ent

bet Weon the variates considered is a measure of their predictive (apabilitv. there

AppelrtoW he s(0.0' at combliniationis amou utdustments whi ch are of (oniparal~e

;Iaitli tv . Just which urnc will he used wilIl he up to the user and the degree of'

Mt~olvetiett desired.

P'redictionis ifr itopetdin't ocuiutenls attempt to loretell what will lake nlace

be tote her at ialiv do aot what the nmgnitd o1 the occurrences will be. The re -

-els i this St tldv ni;v lt )If us(. in the iollokvittg situations:

I. 1l he t111I ilu cet~ ti prton.-s into the polar iooospltere of' the earthI causes

in i'i- lbsorputti )I III, rodio signals that p)ass- touigh tlhe p~olar ionosphere.

I Iisabs-otio ti. krtvmri as polar cap arhsotption or PCA. IhIv toagynitude or thet

its riotmi oil hl. 1"lItu' oi th, 1t) Ml\ pttlon ftltx hr 11i( Followioc (.xpIess.-i~t1

a z 1A2. 0 Km P is the wItoi Ilix it) prlots ((it2 s stetI and A- is tie(

I' I. ~i ~l.. . . ,j 'rtI .iutst romo, 6i. 1.. (M,78) A tiatalo2! or Prtotott I.:rents8- li?. Il'i 2lL Nmn -(lass ical Solar Rarlio hrIS&t,

UP --7 - -.-~IT, -IT7VT30- iVT iWh7f)hh a1 1 nI q4 6111 1T11 Pa psI1)e I-s 6832.

K2. The increased incidence of energetic protons into the polar ionosphere

causes the height of tile D-region to be lowered by changing the D-region

ionization structure. It has been shown that during a proton flux increase, the

height of the south polar D-region tracks the variations in tile proton flux reasonably

well. 25 Changes in the 1)-region height influence the propagation of VLF radio

signals across the polar caps causing variations in their phase and amplitude~26structure, which impacts their use for radio navigation purposes. The integrated

solar radio flux should make it possible to make an earlier determination of the

maxim unt D-region height change to be expected. from which the disruption of the

trans-polar V LF propagation may be anticipated.

3. One requisite of a prediction process is that it provide the prediction as far

ahead of its occurrence as possible. Smart and Shea, 27 suggested a method of

predicting the end of a proton flux increase by extrapolating the end time of the in-

crease after the maximum of the increase has occurred. The process samples he

proton flux levels at a succession of times to obtain the time character of the flux

drop-off and then projects the time when the flux increase will be reduced to zero.

The same procedure can be applied to the radio flux increase being observed, to

predict when the increase returns to zero or to some predetermined level. With

that time and the sense of the functional nature of the radio flux decrease, the

time-integration of the radio fluxes can be accomplished. With these integrations,

the expected peak energetic proton flux can be predicted. [his prediction mav nit

be the best possible because tile position of tile flare on the face of the sun nrax not

be known, but it will at least give a first approxim ati in f the proton flux increase

1o be expected. IF'r a radio burst of clt parativelv sh)rt duratiion, the time advan-

!aue of the prediction may not be great, hut for radio events with longer decav times,

rod for long duration events of the "nonclassical" Itx'pe 'onsidered b Castelli and

Tarristcom2 tile timeliness of the neak proton flux prediction 11ax he ettharicel lv

.as much as an totr or more

25. letni s, \%,. 1. ( 1978) Polar l)-reaion ,lectron densitv prolilt's during a solarproton event. Radio SdJ. 13(No. 5):8.53.

ow 0

26. W\ sterlund, S. , Reder, I. 1. , indAbon. ,C. II 9w)) ofts of polar 'api absorp-tion events on V l\ ' t ransnissions. Planet. Soa, c Sti. 17(No. 7):1:129.

____- L4W27. Sruart, I). 1'., and Shla, M.. A. ( 1117:3) Prediction f the t'trd )I solair iroton

eve.t;,s, Proc. l5th Plenary Alectirg )f COSPAU. .\llril, Spain,10-1. MaV T -7- Space lies. NI II T, eitie verIlle. NIcriri.

References

t

1. Campbell, WIf. et at (1979)) Solar-Terrestrial Predictions Proceedings, Vol. 11,Working Group Reports and Reviews, International Solar -Terrestrial Pre-dictions Proceedings and Workshop, Boulder, Colorado, 23-27 April 1979.

2. (astelli, J. 1P., Aarons, J. . and Michael, G. (1967) Flux density measurementsif' radio bursts of' proton producing flares and non-proton flares, J. Geophys.

,is Vol. 72(No. 2l):5491, ( 19 69 ) Radio Burst S p~~~~a~TFte Short Tie'S Prediction of Solar Proton Events, Proc. of the Symposiumit' Ionospheric Forecasting, Proc. 49, NATO Advisory Group for Aerospace

learid I )UVElopr(mint, C rev H-ocks Houtel,. St. I ovite, Canada, 2-5 Septem be r

:3. C~asielli. .t. P. ( 1968) Observation and Forecasting of Solar Proton Events,*\'CIL-68-0104, AFP660347. -' ....

4. I Sl iv. I). 1K. ( 1964) I'ola r Cap Abso rption, Planet. Space Sri. 12:495.

b"A

. lav, [H. D)., anld Adarrs, C.kW. C 196!t) Ilirieter rreasurrerments, solar proton

rrrrcrsrties and radliatiron dose r-ates, Plarnet. Space Sri. 17:131:3.6i. I'.etia I. A. .,and laozeririti. 1 .1I. (197 1) Eqruatorial and precipitating solar

rotlons in tire rii.rgrosphere: 2. Iliorieter orbservat ions,.. (ieophvs. lies.

O 1. SalleI, H.aser, .A . ,7Strosrio, M. A.r , al Yates, c . K. (1977) "1i rilIIin lay reptir Riprs adetwe vie poar ca; rooactr asoipio ard solar- rroto r-s,

I(:rlrir Sr-. 12:779i.

' akstlri, 1I. , ard arron, .. (1979) lredicltir o s(olar fla. proeto msreslrrr frrom radi burst data, 1i. (norlivs. ar .s. llNr. Its.

KSra ti. It. l. (196.7)0) lire Ise ola Ha Bursts Iredictors )I Itraran

* Lv. nSht Al g nl d. ics, o V )fN U Sola ProtnW Tvens,~o Pr N rtmher 2ymrres1i70

I. Newll., I1P. (1972) I Pekr atirL' on Fal 'r tirx cii' \ P rt r rnts

l. { iirg I.si -1)4 Ii ar trepat77 Absrion Plarst.V~r Ipc t Su4 i. X1f1495. r

7 i(N L __' 4.1

References

11. ClIiver, E. IV. (197 6) Parient -Flare Emission at 2. 8 Ghz as a Predictor of thePeak Absorption of' Polar Cap Events, NELC/TR -20 lb. Naval ElectronicsLaboratory Center, San [Diego, California 92152.

12. Akini'van, S. T. , Fonrichev, V. V. , and Chertok, 1. M. (1978) Estimates of theintensity )! solar protons from the integral parameters of microwave radiobursts. Geontagnet. Aeron. 18(No. 4):395.

1H. C roo)m, 1). 1 _ ( 197 1) Forecastingthe intensity of solar proton events from thelime characteristics of' solar microwave bursts, Solar Phys. 19(No. 1):17 1.

14. H3ak-,hi, P1. , andi Harron, IV. (1978) Prediction of the Ptoton Flux MagnitudesIromr Hadjo) Burst D~ata, A ECH .- T R-78 -0100, AD A05707 1.

15. Hoarton, I\. It. , and Hakshi, P1. (1979) Application of Integrated Radio BurstFluxes to the Prediction of Solar Energetic Proton Flux Increases, Solar-T errte strial Predictions l'roceeding,.s, Vol. Ill: Solar Activity Predictions,Internitionil Solar-Terrestrial Prediction Proceedings and Workshop.Boulder, Colorado, 2:3-27 April 1979.

16. V an iflleheke, 1M. A. I. , Ala Sung. L. S. , and McDonald, F. B. (1975) The varia-tio)n o)f solar proton energy spectra and size distribution with heliolongitude,Solar Phys. 4 l(No. 1): 189.

17. Akin'van, S. T. . Fomichev, V. V. , and Chertok, 1. M. ( 1977) Determination ofthe parameters of solar protons in the neighborhood of the earth from radiobursts : I. Intensity Function. Geomagnet. Aeron. 17(No. 1):5.

18. Reirthard, K. , and Wihberenz, G. (1974) Propagation of flare protons in thesolar atmosphere. Solar Phys. 36(No. 2):473.

V1. Notte. 1. T. , andl lloelol', E.. C. (1973) Large scale structure of the interplanetarynmedium,. 1: Hfigh coronal source longitude of the quiet-time solar wind,Solar lhyvs . 33(No. 0):24 1.

20. Smart, ). I., aind Shea, Al.A. (197 9) PPS76 - A Computerized "Event Mode,"Sola r Proton Fo reca sting Technique, Solar-Terrestrial Predictions Proceed-ings, V'ol. 1, P'rediction Group Reports, International Solar-Terrestrial Pre-dictions Proceedings and Workshop, Boulder, Colorado, 23-27 April 1979.

21. KinL! . If . (1977) Inte rplanieta ry Mledium [Data Book - Appendix NSSDC/WDC-A -

H -S77 -0-1a, National Space Sci. D~ata Center, National Aeronautics and SpaceAdmttinistration. Gorddard Space ["light Center, G reenbelt, Maryland;Inter'planetarv Nledium D~ata Book, Suppl. 1, 1975-1978, NSSDC/WDC-A-H-S 79) 08, I)ec. 1979. Nat. Space Sri D~ata Center, Nat. Aeronautics and SpaceAdmitn. , UIlard Spa( e Fliidht ('enter, Gireenbelt, Md.

4 22., ( osl-ing. . 1 (1976) 'I1 rans ietrt Phenomena in the Solar Atmosphere and SolarIA ind. Phyisics oft Solar Planetar'y Knvironment, Vol. 1, Proc. of InternationalSvinrpsut Wr olrteestriat Physics, Boulder, Colorado, 17 -18 June 1976.

23. Spiegel, ',%. R. (196 It '1lreorv andi l'roblenis of Statistics, Schaum' s OutlineSeties. Schauitr [flub. (>1. . New York, N. Y.

2-1. (as tell i .1. 1'. , atnd morst toni G . L. (1978) A Catalog of Proton Events966- 1967, Having Non-Classical Solar Radio Burst Specttra. AFGL-TR3-78-

1TI21, AD) A0bU816. Environmental Iles. Papers 632, Hanscom AF13, Mlassa-

25. Hlelmis. A.) (1971) Polar 1)-region electron density profiles duting a solarproton iwverit Radio Sci. 13(No. 5):853.

A

C References

26. eItLr I [II, S., lt ,CrC, F. H., and Aborn, C. r1969) Iffects of poIar cap absorp-tifn t- -elt m \ I.I rinsmissions, Planet. Space Sci. 17(No. 7):1:32).

27. Sia ', I). 1'. , a'il Sh a, I M .-A. (197:) Prediction 9f tho end 91f solar' protonvenr, .t. P'o, . 151h 'lena 'v Meeting ,)f COSPAH, Madrid, Spain,

10-1.1 MlavT77 -Zpac' Hes. XII, \ol. 2. Akad(.niie-vrlag. Berlin.

32

6t

I