Magnetic Cloud Field Intensities and Solar Wind Velocities

GEOPHYSICAL RESEARCH LETTERS, VOL. 25, NO. 7, PAGES 963-966, APRIL 1, 1998

Magnetic cloud field intensitiesand solar wind velocities

W. D. Gonzalez, A. L. (",Idade Go•zalez, A. Da.1La,go Institute Nacional de PesquisasEspaciais, She Jos6dos Campos, SP, Brasil

B. T. Tsurutani, .1. K. Arballo, G. K. La,klfina,B. Butt, C. M. SpacePla.sma Physics, Jet PropulsionLaboratory, C, Mifornia Institute of Technology,Pasadena,

S.-T. Wu (',enterfor SpacePlasma and AeronomicResearch, and Departmentof Mechanicaland Aerospace Engineering,The 11niversityof Alabamain Huntsville,Huntsville

Abstract. For the sets of magnetic clo•ds studied in this work [Dungcy,1961 ]. In the aboveexpression, v is the so- we have shown the existence of a relationship between lar wind velocity and Bs is the southward component their peak magneticfield strength and peak velocity of the interplanetary magnetic field (IMF). Gonzalez values, with a clear tendencythat, cloudswhich move and Tsurutani [1987]have establishedempirically that at,higher speeds also possess higher core magnetic field the interplanetary electric field must be greater than 5 strengths.This resultsuggests a possibleintrinsic prop- mV/m for longerthan 3 hoursto createa Dst _• --100 erty of magneticclouds and alsoimplies a geophysical nT magnetic storm. This correspondsto a southward consequence.The relativelylow field strengthsat low field component larger than 12.5 nT for a solar wind velocitiesis pres•mablythe causeof the lackof intense speedof • 400 km/s. stormsduring low speede. jecta. There is alsoan indi- Although the positive correlation })el,weenfast (',MEs cation that, this type of behavior is peculiar for mag- and magnetic storms have been stressedand is reason- netic clouds, whereas other types of non cloud-driver ably well understood, little attention has been paid to gasevents do not,seem to showa similarrelationship, the opposite question, why don't, slow CMEs lead to at least,for the data studied in this paper. We suggest rnagnetic storms? Since this question involves the role that, a field/speedrelationship for magneticclouds, as of the magnetic field of the CME, it, leads to a more that obtained in our presentstudy, could be associated generalquestion, namely how are the CME's speedand with the cloud release and acceleration mechanism a.t magnetic field related to each other? These are the the sun. questionswe wish l,o addressin this letter. Sincefor magneticclouds the total field tyically has a If the speed of the solar ejecta is less than the up- substantialsouthward component, B•, our resultsin,ply stream slow solar wind plus the magnetosonicwave's that the interplanetary dawn-duskelectric field, given phase velocity, a shock will not forin at, the leading (an- by v x Bs (wherev is the cloud'svelocity), is enhanced tisunward) edge of the e;jecta, and there will not be by both factors. Therefore, the consequentmagneto- compressedsheath fields. ttowever,one might ask "why sphericenergization (that is governedby this electric can't the ejecta t,hemselves have fields intense enough field) becomesmore efficientfor the occurrenceof mag- to create magnetic storms with intensities Dst < -100 netic storms. nT7". A•nong the studied relationshipof e.jecta with magneticstorms, practically all intensest, orms were as- socia.ted with magneticclouds [ e.g. Burlaga ctal., Introduction 1987; Tsurutani et al., 1988] that followedshocks, although Tsurutani ctal. [1988]claimed that the intense It has been well establishedthat major (D.,t _( -100 storm of September18 1979, wasassociated with a mag- nT) magnetic storms are associatedwith fast, Coronal netic cloud without a shock (with an average speed Mass Ejections((•MEs) comingfrom the sun from both around 370 kin/s). This latter type of associationap- duringsolar maximum [Tsurulani el al., 1988, Gosling pears to be uncommon. et al. 1991, Gonzalezet al., 1994] and the descending phaseof the solar cycle [Tsur•tani ctal., 1995]. The energytransfer mechanisnifroni the solar wind to the Method of Data Analysis. magnetosphereappears to be magneticreconnection between the interplanetary magneticfield and the earth's To address this issue first we use magnetic cloud magneticfield. The interplanetary dawn-duskelectric events.Magnetic clouds [e.g. Burla•a, 1995]are a type field, which is given by v x B•, governs this process of ejecta that have a mesoscaledimension of about 0.25 AU at 1 AU and crossesthe spacecrafttypically in about Copyright1998 by theAmerican Geophysical Union. 24 hours. Magnetic cloudshave a large and smooth rotation in the field's direction, an enhancedfield strength Papernumber 98GL00703. and low proton temperature and/• values. It is thought 0094-8534/98/98GL-00703505.00 [Farragiae! al., 1997and references therein] that clouds

963 964 GONZALEZ ET AL.' MAGNETIC CLOUD FIELDS AND VELOCITIES are giant flux ropes formed by field aligned currents. Second Set, of Clouds (1979) The B•, componentof the cloud has typically an ampIf i ß ! ! ! ß tude that represents a substantial [raction of the total 25 IMF intensity. For example, for the intense and super- intensestorms studied by Tsurutaniel, al. [1988]and Tsurutani et al. [1992],the magneticclouds respon- 20 sible for about half of the storm events had their B., fields with intensities of 70% or more than that of the m 15- correspondingtotal IMF intensities. This association can explain why magnetic cloudswith strong magnetic fields have typically strong B, componentsand there- 10 fore cause intense storms. We have chosento examinefirst previouslypublished magnetic cloud eventsbecause their identificationexists õ in the literature and the readerhas the opportunityto examinethe eventsin detail. We use(1) the five cloud 300 350 400 450 500 550 600 eventspublished and illustrated by Klein and Burlaga Peak v (kin/s) [1982]; (2) the two cloudevents published and illustrated by Burlagaet al. [1987],•sing data.[rotl• t,he Figure 2. Scatterplot ForBr,•. vers,sv•,•,a•- for the secondset of 13 •agnetic clo;•(!s,as observedby the IMP8/ISEE-3 satellites;(3) six cloud eventsreported ISEE-3 satellite in 1979. These evc•l,s i•volved clouds by Tsurutaniet al. [1988]and Tsurutaniel al. [1992], that causedintense and •o(!erate magneticstorn•s. usingISEE-3 data;(4) threecloud events discussed by Farrugiaet al. [1997]who refer to previousstudies that useddata from the IMP8/ISEE-3 satellites;and (5) the one event,reported by Burlagaet al. [1996],using data satellite, and identified the driver gas events using a from the WIND satellite. All these 17 events were ob- combined criteria from those discussedby Zwickl et al. served at • 1 AU and the criteria to select the events, [98.q] Ts,vttani at. [988]. The combinedcri- that are cormnon in all these cases, are: a large ro- teria involved' a smooth ]nagneticfield (lack o[ Alfvdn tation in the ejecta field's direction, an enhancedfield wavesand discontinuities),high magneticfield strength, strength (typically > 10 nT), a low proton te•nperature and a low proton temperature. For this study we have and a duration of about, 24 hours for the ejecta to cross used5 to 15 minute averagesof the ISEE-3 data to iden- the spacecraft. tify the events and higher t,ime resolutionto checkthe Because most of these events were studied in asso- smoothness of the field. These events were dividcd in ciation with intense magnetic storms, except,the five two subsets, one including only clear magnetic clouds, eventsreported by Klein and Burlaga[1982], who se- using the selection criteria mentioned in the previous lected clouds without any association with magnetic paragraph (with field rotation anglesthat for this data, storms, we have looked for another (and independent) set lie in the domain of abo]•t 200 to 320 degrees),and set of magnetic cloudsthat would not have an (a pri- the other subset was formed with the remaining driver ori) associationwith magneticstorms. We obtained gas-non cloud events. For the first s•bsot we obtained this secondset of clotadsin the followingway. We took 13 events and for the second one 24 events. the whole year of 1979, for which a fifil set,of plaslna As for the first, set, of clouds, we also performed the and magnetic field data was recordedby the ISEE-3 same study for the peak values of the solar wind speed and the magnetic field within the intervalsof these two subsets of events. First Set, of Clouds It is interesting to point, out that the secondset, 40 ' i ' i ß i i of cloud events, even though they were selected without looking to their relationship to magnetic storms,

35 they were later seen to still be associatedwith magnetic storms of at. least,moderate intensities (-100 nT _< Dst <_-50 nT). One may think that this could be due 30 to the fact that the year 1979 was near solar maxi]num, but. it, has been shown that, the solar cycle distribution 25 of storms does not necessarilyfollow the sunspot solar cycledistribution [e.g. Gonzalezctal., 1994]. a0 On the other hand, it is also possiblethat magnetic clouds(following shocks)are ]nest of t.!•cl,i•es associated with intense storms or at least with st.orn•s of 15 moderate intensity, becausethey tend to involvefairly

10 , I I • I , I , large valuesof B.• [e.g. Bnrlagact al., 1987' Tsurutani 3OO 400 ,500 600 700 800 c! al., 1988]. Peak v (kin/s) Results. Figure 1. Scatter[)lot, k)r Bl,eat: versus •.'•,•a•: for the first set of 17 magnetic clouds, taken from several ref- The magneticcloud intensity versusspeed for e.ach of erences(see text), involving clouds that ca,•se(t the 17 eventsof the first set,of cloudsis shownin Figure intense and very intense •nagnet,icst, orn•s. 1, iu a scatter-plot forn•at (Prcli•]•inary results related GONZALEZ ET AL.- MAGNETIC CLOUD FIELDS AND VELOCITIES 965

Combined Set of Clouds Driver Gas - Non Clouds 4o ß i ß i ß i ß i ß i ß i ß I ß ! ß i

m 20

ß 0 ß I • I • I • I • i • I • I • I • I 300 400 500 600 700 800 300 400 1500 600 700 Peak v (kin/s) Peak v (kin/s) Figure 3. Scatter plot for the coml,ination of tlw two Figure 4. Driver gas eventsthat apparentlydid not independentsets of cloudsshown in Figures1 and 2. involveclouds in 1979(see text), as measuredby the ISEE-3 satellite.

to this figurewere briefly mentioned by Tsurutanictal. velocity is presumably the cause of the lack of intense [1998]).A linearregression fit is addedto theplot,. The storms d•ring low speedejecta. correlationcoefficient is 0.71 and the linearregression There is also an indicationthat this type of behav- line gives, ior is peculiarfor magneticclouds, whereas other types of non cloud-drivergas eventsdo not seemto show a Bp,•.(nT) = 0.047,,•,•(km/s) - 1.l similar relationship. Due to the B-v relationshipfor magneticclouds that Figure 2 shows a similar plot, for the secondset we presentin this paper, a fastercloud implies a larger of clouds,namely for the 1979 driver gas-clearcloud amplitude of B (which typically has a substantial B• events, and Figure 3 refers to the combineddata set,of component). Thus, since B• and v are related to one thoseplotted in FiguresI and 2, involving30 clouds. another, the magnetosphericenergy transfer, which is The correlation coefficients obtained from the linear fits governedby v x Bs [e.g. Gonzalezet al., 1994],is en- in Figures 2 and 3 are, respectively,0.73 and 0.7.5. hanced by bothfactors and therefore becomesmore effi- From Figure 3, we note that, at cloud velocities of cient for the occurrenceof magneticstorms. This rela: v m 400 km/s, the field magnitudescan be 1,5to 20 nT, tionshipexcludes the possibilitythat a magneticcloud which could have a B., componentco•npara.ble to the with a high speed and a negligibleor absent B., field valueexpected from the Gonzalezand Tsurutani[1987] could lead to the developmentof an intensestorm. To criteria, discussedin [he Introduction. It, is also inter- our knowledge no one has observedsuch a case. This esfing [o note t,ha[ the largely documentedmagnetic lack of observationalso reinforces the B-v relationship cloud event of Jan 10, 1997, with peak valuesof about, reported in our paper. 1,5nT and 480 km/s, fits fairly well the trend shownin At.this time, the physicalcauses of the reportedrela- Figure 3. tionship betweenthe magneticfield and plasmaspeed Figure4 refersto the ISEE-3 subsetof drivergas-non of the magnetic cloudsare uncertain. Compressionof cloudevents. One cansee that this plot is largelyscat- the cloud is certainly occurring. Thus, it is possible tered, without any cleartrend for a relationshipbetween that in some cases the field increases can be accounted the peak valuesof the magneticfield and the solarwind for by such a.neffect. speed. One can speculatethat this set of events could Another possibilityis that this relationshipmay be still include cloudsthat were crossedby the satellite far associated with the CME release and acceleration mech- from their center, for which a rotation in the field was much smaller. In such a case much lower B fields are anismsat the sun. For example,a recentself-consistent numericalMHD simulationstudy [Wu e! al., 1997] expectedfor a similar v value of the cloud. It is pos- showsthat the radial velocity of the center of the flux sible that other driver gas structurescould be involved rope (i.e. the centerof the magneticcloud), erupting in Figure 4, and deservea detailed investigation. from a helmetstreamer, is proportionalto the strength of the azimuthal componentof the magneticfield, in Discussion and Conclusions. a way consistentwith the magnetic cloud B- v re]a- tionshipreported in this paper. A detailed parametric For the sets of magnetic cloudsstudied in this work study of this kind of simulationalwork is currentlyun- we have shownthe existenceof a relationshipbetween der way. their peak magneticfield strength and peak velocity values,that may suggesta possibleintrinsic property Acknowledgments. Portions of this work were sup- of magneticclouds and also imply a geophysicalcon- ported by the Fundo N acional de Desenvolvimento Cient[fico sequence. There is a clear tendency that cloudswhich e Tecno16gicoof Brazil and by the National Aeronautics and moveat higherspeeds also possess higher core magnetic SpaceAdministration/ Jet PropulsionLaboratory, Califor- fieldstrengths. The relativelylow field strength at a low nia Institute of Technology. 966 GONZALEZ ET AL.: MAGNETIC CLOUD FIELDS AND VELOCITIES

References T. Lee , Great magnetic storms. Geophys. Res. Lett., 19, 73 , 1992. Burlaga, L. F., Interplanetary Magnetohydrodynamics,Ox- Tsurutani, B. T., W. D. Gonzalez, A. L. C. Gonzalez, F. ford University Press, New York, 1995. Tang, J. K. Arba!1o,and M. Okada,Interplanetary origin Burlaga, L. F., K. W. Behannon, and L. W. Klein, Com- of geomagneticactivity in the decliningphase of the solar pound streams, magnetic clouds, and lnajor geomagnetic cycle. J. Geophys.Res., 100 (All), 21717, 1995. storms, J. Geophys. Res., 92, 5725, 1987. Tsurutani, B. T., J. K. Arballo, Y. Kamide, W. D. Gonzalez, and R. P. Lepping, Interplanetary causesof great. and Burlaga, L. F., R. P. Lepping, W. Mish, K. W. Ogilvie, A. superintensemagnetic storms, Physicsand (7hemistryof Szabo, A. J. Lazarus, and J. T. Steinberg, A magnetic the Earth, in press, 1998. cloud observed by Wind on October 18--20, 1995, NA,qA Wu, S. T., W. P. Guo, and M. Dryer, Dynamical Evo- preprint, G,qFC, February, 1996. lution of a Coronal Streamer-Flux Rope System: II. A Dungey, J. W., Int.erplanetary magnetic field and the auroral Self-ConsistentNon-Planar Magnetohydrodyna,nicSim- zones, Phys. Rev. Left., 6, 47, 1961. ulation, Solar Phys., 170, 265, 1997. Farrugia, C. J., L. F. Burlaga, and R. P. Lepping, Magnetic Zwickl, R. D., J. R. Asbridge,S. J. Bame,W. C,.Feldman, J. clouds and the quiet-storm effect. at earth, in Magnetic T. Gosling,and E. J. Smith, Plasmaproperties of driver Storms, edited by B. T. Tsurutani, W. D. GonzMez, Y. gas followinginterplanetary shocks observed by ISEE-3, Kamide and J. K. Arba!1o, Amer. Geophys. lYnion Press, in Solar Wind Five, NASA Conf. Publ., CP. 2280, 711 Washington D. (',., Geophys. Mon. Set., 98, 91, ]997. pp., 1983. Gonzalez, W. D., and B.T. Tsurutani, ('•riteria of interplanetary parameters causing intense magnetic storms (Dst < -100nT). Planetary ,qpaceScience, .•5, 1101, W. D. Gonzalez, A. L. Clfia de Gonzalez, A. dal Lago, 1987. Instituto de PesquisasEspaciais, 12201-970, S•.o .1os•.dos Gonzalez, W. D., J. A..loselyn, Y. Kamide, H. W. Kroehl, C,ampos, CP 515, SP, Brasil. (e-mail: gonzalez½$dge.inpe.br; G. Rostoker, B. T. Tsurutani, and V. M. Vasyliunas, alicia•dge.inpe.br; dallago(•dge.inpe.br) What is a geo•nagneticstor•n?", Review Paper. J. Geo- B. T. Tsurutani, J. K. Arballo, G. K. Lakhina, B. phys. Res.,, 99 (A4), 5771, 1994. Butt, C. M. Ho, Space Plasma Physics, Jet Propulsion Laboratory, California Institute of Technology,Pasadena, Gosling, .1. T., D. J. McComas, J. L. Phillips, and S. J. California, 91109. (e-mail: BTSURUTANI•jplsp2.jpl.nasa. Bame, Geomagnetic activity associatedwith earth pas- sage of interplanetary shock disturbances and coronal gov; JARBALLO•_ jplsp2.jpl.nasa.gov;LAKt{INA½•jplsp2. mass ejections. J. Geophys. Res., 96, 7831, 1991. jpl.nasa.gov;BBUTI•_jplsp2.jpl.nasa.gov; [email protected]. na•a.gov) Klein, L. W., and L. F. Burlaga, Interplanetarymagnetic S.-T. Wu, Center for Space Plasma and Aeronomic clouds at 1 AU, ,1. Geophys. Res., 87, 61:3,1982. Research, and Department of Mechanical and Aerospace Tsurutani, B. T., W. D. Gonzalez, W. D., F. Tang, S. I. Engineering, The University of Alabama in Huntsville, Akasofu, and E. J. Smith, Solar wind southward B z fea- Huntsville, Alabama 35899. (e-mail: [email protected]) tures responsiblefor major magnetic storms of ] 978-1979. J. Geophys.Res., 93 (AS), 8519, 1988. (ReceivedNovember 10, 1997; revisedDecember 26, 1997; Tsurutani, B. T, W. D. Gonzalez, W. D., F. Tang, and Y. accepted February 10, 1998.)