JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 99, NO. AI0, PAGES 19,213-19,227, OCTOBER 1, 1994

Neutral properties: Advance warning of major geomagneticstorms

Michael A. Gruntman Departmentof AerospaceEngineering, University of SouthernCalifornia, Los Angeles

Abstract. Coronal massejections (CMEs) have been identifiedas a trigger for large geomagneticstorms. A clever idea to provideadvance warning of a high-spee•CME approachingthe was recentlyproposed by Hsieh et al. (1992)' ejectedsolar matter decelerateson its way from the Sun to Earth, energeticneutral (ENAs) are formed in the CME plasmadue to recombination,ENAs passthe deceleratingCME plasmaand arrive first at the Earth. We evaluate the idea to use ENAs for advance detection of the high-spee• Earth-approachingCMEs and considerthe processesinvolved. Charge exchangebetween solar wind and interplanetaryneutral atoms contribute effectively to ENA production. Characteristicsof neutralgas within Earth's are updatedfor both neutralatoms of interstellarorigin and from outgassingfrom interplanetarydust. Computersimulation of CME-producedENAs showsthat the total flux of CME-produced ENAs is slightlysmaller than the intensityof the quiescentneutral solar wind (NSW) while significantlyhigher ENA energiesmake them easily distinguishablefrom NSW atoms. Arrival of the CME-produced ENAs is expected3-4 hours before the start of a large geomagneticstorm, which providesa basisfor advancestorm warning and predic- tion of storm magnitude. The progressin developmentof ENA measurementtechnique suggeststhat both the quiescentNSW and CME-producedENAs can be reliably mea- sured.

1. Introduction The phenomenonof CME is not completelym•lerstood, Coronal massejections (CME) have been identifiedas a and no easily observablephenomenon at the Sun, suchas trigger for large nonrecurrentgeomagnetic storms [e.g., solar flares, has yet been identifiedthat allows a reliable Gosling et al., 1991; Gonzalez and Tsurutani, 1992; advancewarning of a large CME approachingthe Earth. Tsurutaniet al., 1992a;Gosling, 1992, 1993a,b; Tsun•tani Obviously,CME patrol can be providedby a spacecraft and Gonzalez,1993]. An ejectionof largequantities of fast somewherebetween the Earth and the Sun. The Lagrangian moving(50 to > 1200km/s) solar material into interplane- point L1 (of the Sun-Earthsystem) is a convenientplace tary spaceoccurs sporadically, and approximately1% to for sucha monitor. However,this point is only 1.4 x 10• 10 % of all solarwind plasmameasurements were madein km from the Earth, allowingonly 30-rainadvance warning CMEs during solar minimumanti maxinmmresl•tively before a CMEs impingementupon the . [e.g., Gosling,1992]. The Earthintercepts annually about CMEs are reliably observedby white light coronagraphs 70 CMEs at solar activity maximumand less than 10 that measuresolar light scattered(Thomson scattering) by CMEs near solar activity mininmm. Only a fractionof plasmanear the Sun [e.g., Hundhausen,1993]. A deploy- interceptedCMEs causesmajor geomagneticstorms, which ment of spacecraftwell aheadand behind the Earth in its also requirethe presenceof an intenselong-duration south- orbit aroundthe Sun wouldallow the detectionand velocity determinationof a CME leaving the solar stirface and wardinterplanetary [e.g., Gosl#tg,1993a, b; Tsurutani and Gonzalez, 1993]. directedtowards the Earth [Goslinget al., 1991]. Optical Major geomagneticstomas create electrical power detectionof such CMEs from a spacecraftnear Earth was outages,disrupt conmmnications,damage and reducethe discussedby Jacksonet al. [1991]. lifetimeof Earth-orbitingsatellites, and posehealth hazards A clever idea to providean advancewarning of a high- to airline passengersat high altitudes[e.g., Tsurutaniet al., speedCME approachingthe Earth was recently suggested by 1992a;Lanzerotti, 1992]. Althougha clear understanding Hsieh eta/. [1992]. They proposedto take advantageof the of the relationbetween the disturbedspace weather condi- fact that ejectedsolar matter decelerates on its way from the Sunto Earth. Someof the plasmaprotons would recombine tionsand adverse effects on technologicalsystems has yet to with electronsproducing energetic neutral atoms (ENAs) be achieved[Lanzerotti, 1992; Tsurutaniet aL, 1992b], an moving with the local velocity. ENAs move advancewarning of largegeomagnetic storms and prediction independentlyof the plasnm,and someof the ENAs would of storm magnitudeare highly desirableand would help eventuallyovertake the leading edge of the decelerating reducestorm-related damage. CME and arrive first at the Earth. Thtksby measuringan Copyright1994 by the AmericanGeophysical Union. ENA flux onecan obtain a precursorsignal of the approach- ing CME beforeits impingementtipon the magnetosphere. Paper number94JA01571. As it was pointedout by D.McComas(private conununi- 0148-0227/94/94JA-01571505.00 cation, 1994), a shockpreceding the bulk of CME plasnm

19,213 19,214 GRUNTMAN: NEUTRAL SOLAR WIND - WARNING wouldreach tho magnotosphoreearlior than tho CME. Wo oxperhnontalapproach to tho CME detection is briofiy shouldnoto horo that (1) not all shocksin intorplanotary discussed. spacoaro accompaniedor producedby CMEs and (2) tho regionbetween tho forward shock and CME is filledby hot 2. Phenomenon of Neutral Solar Wind fast-movingplasma which would also gonorato fast ENAs. Considorationof comploxoffects of tho forward shocks Interactionsbetween charged and neutral particlesare precedingCMEs is boyondtho scopeof this work andwo commonphonomona in spacoplasmas. In tho solar wind assumohoro that geomagnoticstorms start whentho CME thesoaro manifestedby tho presencoof a solarwind neutral plasmahnpinges on tho magnetosphoto. componont(NSW). It was speculatedlong ago that a largo It shouldbo mado clear that a significantfraction of numberof noutralatoms could bo presontin tho solarwind CMEs do not trodergoany decoleration,and manyundergo as transiontsduo to ojectionof solarmattor in violontovonts accoloration,through thoir intoractionwith tho solarwind. [Akasofu, 1964a, bl. Noutral hydrogonatoms in solar Tho largoand major geo•nagnoticstorm catogories aro of prominoncesaro obsorvedoptically, and it is arguedthat mostintorest for advancostorm warning. It is suspectedthat thoy may reach 1 AU [Illing a•ut Hild•r, 1994]. Tho "thoinitial speedof a CME relativoto tho ambientsolar probability to reach I AU without ionization depends wind aheadis probablytho mosthnportant factor in detor- stronglyon hydrogonatom velocity, solar EUV , miningif an earthwarddirected ovont will bo offectivoin and plasma tomperamroand it does not oxceed oxcitinga largogeomagnotic disturbanco" [Gosl#•g, 1993a, 0.01. The quostionwhothor a substantialamount of hydro- p.2644]. Thoroforowo lhnit our considorationin tiffswork gon atoms is presontat 1 AU is open. In any caso, if to high-speedCMEs only. noutralsojected from the Sun reachtho Earth, thoy would Tho estimatosof tho ENA flux obtainedby Hsieh et al. providoan ENA flux in additionto tho flux considoredin [1992] woro incomplotoand did not allow tho feasibility this paper. In this casotho advancowarning teclmiquo ovaluationof tho proposedteclmiquo for soyoralreasons. would be ovon moro officiont. First, tho noutralcomponont is pennanontlyprosont in tho A small fraction of plasmaparticles is noutralovon at "rogular,"quiescent solar wind - tho noutralsolar wind high-tomporaturoconditions in tho solar corona. For (NSW). ThoCME-produced ENA flux hasto bocompared examplo,neutral atomsaro presentin tho oxpmxlingsolar with tho stationaryNSW sincotho omhancomontof ENA coronaduo to recombination,tho process invoked by Hsieh flux, and notjust its appearanco,as suggostedby Hsiehet et al. [1992]. Thoro is, howovor,anothor process that al. [1992], wouldboa precursorof thoapproaclfing CME. contnlmtesto a buildup of tho NSW: chargooxchango (Wo will fi•rthor refor to such CME-producedENAs as collisionsbotween solar wind ions and noutral particles CMF_./ENAsto distinguishthom from tho ENA flux in tho filling intorplanetaryspaco. Tho realizationof tho fact that stationarysolar wind, i.o., NSW). Second,both tho quios- intorplanetaryspaco is filled with neutralatoms led to tho cont NSW and CME/ENAs aro formed by contributions dovolopmontof a concoptof a permanontlyoxisting NSW. from sovoralprocesses and tho reco•nbination in tho oxpmxl- Thotwo majorsources of noutralgas insido tho Earth's orbit ing solarplasma is only onosourco of ENAs. It is notyet aro intorstoHargas (ISG) penetratingtho solar systomfrom clearhow othorsources compare with tho recombinationin tho local intorstoHarmedium (LISM) and neutralizationof tho CME plasma: tho recombinationprovides a nilnor solar wind ions by intorplanotarydust. ENAs born in a contnt>utionto tho quiescontNSW [o.g., Holzer, 1977], togion of tho holiosphoricintorfaco would providoonly a buthighor nmnber donsities and lowor electron tomperatures small contribution. Outgassingof planetsproduces local typicalfor CME plasmamay (or may not) changotho emhancomontsof neutral particlo ntm•ber donsity and can bo relativeimportanco of difforentprocesses. Finally, tho disregardedin tho framo of tho presontwork. modolused by Hsiehet al. [1992]did nottako into account Tho NSW atoms, born in recombinationand chargoox- sphoricaldivergonco of tho CMFJENA flux wlfichwould- changoprocesses, moro in an antistmwarddirection with tho lowortheir ENA flux estinmtesby 1-2 ordorsof magnimdo solar wind velocity. Tho NSW atomsaro lost in collisions at 1 AU from tho Stm. with tho solar wind olectronsand ions and duo to photo- Thogoal of thiswork is to followtho or/ginal suggestion ionization by tho solaroxtremo ultraviolet (EUV) radiation. by Hsiehet al. [1992]to useENAs for advancodetection A relativolyhigh rolocity of tho solar wind (-500 knds) of high-speedCMEs approachingtho Earth and considor in allowsmany ENAs to travollargo distances without ioniza- detailtho processes involved. This would allow tho feasibil- tion and ovonmallyleavo tho holiosphoro. ity ovaluationof tho proposedtechniquo for advancogeo- Tho NSW flux composition,intonsity, and enorgy magneticstorm warning and prediction of storm•nagnitudo distribution dopend upon tho solar wind charactoristicsand and lead to tho detorminationof requiromontsfor corro- distribution of neutral particles insido tho Earth's orbit spendingspacecraft instmmontation. We shouldalso noto arotmd tho Stm. A dovolopmontof tho modol of solar that ENA measuremontsin tho solar wind is a natural way systompenetration by ISG allowed OhOto calculatotho of experimontallystudying tho ovolutionof CMEs as thoy oxpectedneutral numberdonsities in intorplanetary propagatooutward from tho Sun. Changoof CME sizo, spaco[Fahr, 1968a]. This modol was used to obtain an speed, plasmatomperature and numberdonsity would NSW flint [Fahr, 1968b; Blum and Fahr, 1970] duo to manifestitsolf in thotomporal variations of thointonsity and chargooxchango between solar wind and intorstollar enorgydistributions of thoCMFJENA flux. hydrogonatoms. Lator it was recognizedthat chargo First, wo will briofly reviow tho concoptof tho NSW, oxchangoof protonswith intorplanetaryholitm• atoms may thonupdato neutral particlo number donsities insido tho providoa significantcontnl)ution to tho NSW in tho region Earth'sorbit and calculatotho stationaryNSW fiux. Thon, within sovoralastronomical units (AU) from the Stm [Gnou- charactoristicsof CMF_./ENAflux is obtainedand finally, an man, 1980]. A computorsimulation of tho NSW angular GRUNTMAN: NEUTRAL SOLAR WIND - GEOMAGNETIC STORM WARNING 19,215 characteristics(NSW isofluxesas observedfrom the moving Directexperimental study of the NSW facedseemingly Earth) wasperformed by Bleszynskiet al. [1992] (tiffswork unsurmountableexperimental difficulties. However, was actuallydone almost 10 yearsago, althoughpublished experimentaltechniques to measurelow-energy (< 10 keV) only recently). V.B.Leonas(private communication, 1987) ENAs havebeen gradually developing since the late 1960s draw attention to the fact that the cross section for simulta- [e.g., Wax and Bernstein,1967; Bernsteinet al., 1969; neoustransfer of two electronsto solarwind alphaparticles Gruntmanand Morozov, 1982; McEntire atui Mitchell, in collisionswith heliumatoms is fairly high. The conse- 1989;McComa, et al., 1991;Hsieh et al., 1991;Gruntman, quence of such a double electron charge transfer is a 1993; Funstenet al., 1993]. An experimentalconcept to nonnegligiblehelium atom component in the NSW, espe- measurethe N SW wasproposed and developed in theearly cially at distancesof several AIJ from the Sun, and a 1980s [Gruntman atut Morozov, 1982; Gruntman atul presenceof doublycharged ions among the solar Leona,, 1983; Gruntmanet al., 1990]. A dedicatedNSW wind pickupions is expected. experimentwas includedin the scientificpayload of the The ISG fills the entire heliosphereon a global scale. SovietRELIKT-2 missionwhich was originallyscheduled Anothersource of neutralatoms, confined to the innerpart for the secondhalf of the 1980s and has not yet been of theEarth's orbit, is interplanetarydust which congregates launched. towardthe Sunclose to the eclipticplane. This lattersource was consideredin detail by Banks [1971]. Solar wind 3. Neutral Atoms Inside Earth's Orbit bombardmentof a dust grain results in a penetrationof hydrogenand helium ions into the surfacelayer to thedepth Dilute neutralgas insidethe Earth's orbit provides of severalhundred angstrems. This layer soon(less than collisionpartners for chargeexchange, which is themajor processin theformation of theNSW. Importantsources of few years)becomes saturated by intruderions, aml outgas- neutralparticles inside the Earth's orbit are the IS(3, out- singof neutralatoms and molecules from the grainmaintains theequilibrium. The estimatesof thisneutral particle source gassingfrom interplanetary dust, and ENAs from a region of the heliosphericinterface. Sincewe are interestedin the sufferfrom a largeuncertainty in dustgrain population and parametersof theNSW reachingthe Earth,we will concen- outgassingprocess details [Banks, 1971; Holzer, 1977; Fahr trateon neutralgas in theecliptic plane only. While number et al., 1981]. Banks [1971] first estimatedthe NSW flux densityof neutralparticles due to dustoutgassing is distrib- due to chargeexchange of the solar wind protonson the uted isotropicallyin the eclipticplane, interstellarand dust-producedneutrals. heliosphericneutrals are highlyanisotropic with a vectorof The neutral componentin the solar wind constitutesa the LISM velocityrelative to the Sunproviding an axisof rather small fraction, ---10'n, of the solar wind at 1 All symmetry. The most comprehensivediscussion of neutral from the Sun (see section4). As the solar wind expands gas parametersinside the Earth's orbit can be found in the further toward the boundariesof the ,this worksby Holzer [1977] and Fahr et al. [ 19811. We will fraction graduallyincreases to 0.1-0.2 at the solar wind updatetheir resultsand presentneutral particle densities terminationshock and plays an importantrole in global withinthe Earth'sorbit as it pertainsto the NSW. processesin the beliesphere.While the solarwind super- sonicplasma flow is believedto be terminatedby theshock, 3.1. Interstellar Gas the NSW atomseasily penetratethe region of the helio- Theinteraction of theSun and the LISM is manifestedby sphericinterface, enter the LISM and interact(via charge thebuild up of a heliosphere[e.g., Dessler,1967; Axford, exchange)with the approachingflow of interstellarplaslua. 1973; Fahr, 1974; Heizer, 1977; Thonu•, 1978; Bertaux, The resultof this interactionmay becomeespecially hnpor- 1984; Baranov, 19901. It is currentlybelieved that the tantif theflow of interstellarplasma is supersonicand forms interstellarwind comesfrom the directionof ecliptic the bow shock(two-shock model [e.g., Baranov, 19901)in longitude252 ø and latitude +7 ø [e.g., Lallementet aœ, from of the heliosphere.In this casethe interstellarplasma 19901. (We will further assumethat the interstellarwind flow is not supposedto •leam• aboutan obstacleahead, that relative velocityis in the eclipticplane.) The Earth is is, the beliesphere,until reachingthe bow shock. However, positionedin the upwinddirection in thebeginning of June the NSW atoms, •sneaking• outsidethe heliosphere,would eachyear. It is convenientto usethe sphericalcoordinate disturbthe interstellarplasma. Hencethe LISM is modified system, 01,0), with the Sunat the originand angle 0 by the NSW evenbeyond the bow shock. The boundaryof measuredfrom the upwinddirection (Figure 1). the regionof the Sun'sinfluence is thuspushed further into Our calculationsare made here on the basis of thewidely interstellarspace and the inferenceof the •pristine• LISM accepted•hot • modelof IS(3penetration in thesolar system. parametersfrom interplanetaryglow observationsbecomes increasinglycomplicated. This NSW effect•on the L!SM was first predictedby Gruntman[1982] and confirmedlater LISM by a detailedcomputer simulation of Barahoyand Malama [19931. The NSW flux hasnever been studied experimentally, and . there are no direct experimentaldata on the NSW. The measurementof pickupions in the solarwind [e.g., Gloeck- let et al., 1993, andreferences therein], some of thembeing anotherproduct of the same processthat gives rise to the Figure1. Motionof thelocal (LISM) NSW, is an independent,though indirect, confirnmtionof relativeto theSun; angle • is measuredfrom the upwind the validityof our understandingof the processesinvolved. direction; Rg = 1 AU is the radiusof the Earth'sorbit. 19,216 GRUNTMAN: NEUTRAL SOLAR WIND - GEOMAGNETIC STORM WARNING

Numberdensity distributions of neutralatoms in interplane- velocityand temperatureof the ISG, V(oo) anti T(oo), tary spacedepend on theundistnrbed (i.e., "atinfinity") ISG wereidentical for bothhydrogen and helium components, a parameters,density, velocity, temperature, and uponsolar naturalapproach for a gas in equilibrium. However, there radiationpressure, and atonficionization rates. The ISG is a growingperception that the hellospheric interface region mainly consistsof hydrogenand heliumatoms, and solar at a distanceof 70-150AU fromthe Sun plays an important radiationpressure is usuallydescribed by a ratio,/z, of the role in filteringneutral atoms and changing their characteris- averageradiational-to-gravitational forces acting on individu- tics. This effect, first pointedout by Wallis[1975], may al atoms. The solar wind is assumedto be a steady, substantiallychange the parmnetersof interstellarhydrogen sphericallysynm•etric radial flow with constantvelocity, withoutaffecting those of interstellarhelium ato•ns [Ripken Vsw(R) = Vsw= const, and with flux and nulnberdensity amt Fahr, 1983; Baranov, 1990; Fahr, 1990; Baranov atul inverselyproportional to the squareof the distancefrom the Malama, 1993]. The physicalmechanism responsible for Sun, thisinteraction is chargeexchange between plasma ions and neutral atoms. A significant difference in cross sections iVsw,,()= determinesthe difference in the magnitudeof the effect betweeninterstellar and helhm• atoms. Recently N,(a) = (aria) (2) performedsimulation of the solar wind interactionwith the LISM, that for the first time self-consistentlyaccounted for F•sw,• = Fsw,t(Re)and N• = Nt(Rœ)are the fitix and ionizedand neutralcomponents, spectacularly demonstrated numberdensity at Re = 1 AU, that is, at the Earth's orbit. the importanceof this effect [Barahoyatut Malanm, 1993]. •e subscriptI studs for solar w•d ions, either H+ or The major implicationof charge exchangein the hello- He+ +; a capital letter N is rescued for number sphericinterface is that velocityand temperature of interstel- densities, wNle a small letter n will denote the number lar hydrogenand helh•mat infmity, as derivedfrom obser- densiWof neutral paaicles. vation of ISG in the inner part of the solar system(< 30 IoNzation of hydrogen•d helium atomsis assumedto be ALl) and basedon the conventionalhot model, correspond due to photoio•zation by the solar EUV radiation •d to different states of the ISG. In case of helium the derived charge exch•ge with the solar w•d ions. Neutral gas parameters,n.,(oo), V.c(oo), and T.,(oo), correspondto wit• 1 AU is essentiallyoptically t• for solar radiation "genuine"interstellar helium paran•etersin the unperturbed •d the probabilityof ex•rienc•g two cons•utive charge LISM while in case of hydrogen,the parmneters, nu(oo), exch•ges by the sine paaicle is negligible. Radial V.(oo) and Tu(oo), actuallycorrespond to thoseof hydrogen deandentesfor ioNzationrates are •.(R) = •.(Re) (Re/R)2 already modified by passage through the heliospheric •d •H,(R)= •H,(R•)(R•/R)2, where •.(a•) •d •H,(R•) interfaceregion. This wasconvincingly continned by recent are io•zation ratesof hydrogen•d heliumatoms at I AU measurementsof the ISG velocity and temperaturein the from the Sun. •e pr•on•t m•hmsln for helium LISM beyond 1000 AU [Lallementet al., 1993; Bertin et io•zation is photoio•zation, wlfile both charge exch•ge al., 1993]. The values obtained, V,(oo) = 26 kdn/s and •d photoio•zation are im•a•t for hydrogen atoms. T.(oo) = 7000 K, are in excellentagreement with interstel- •erefore the hydrogenio•zation rateis •H = •" + •x, lar helium parametersderived frown direct helium flux where•" is thephotoio•zation rate, • = F•,H+ qH+,H measurementsby [Witte et aL, 1993], but they is the io•zation rate due to chargeexch•ge with solarwh•d significantlydisagree with the interstellarhydrogen parame- protons•r•on•t ch•el), •1 qH+,H is the charge ters derived from "conventional" interplanetary glow exc•ge crosss•tion betw•n protons•1 hydrogenatoms. experiments[e.g., Bertauxet al., 1985;Ajello et al., 1987]. Suchassumptions are widelyaccept• for studyof •te•l•- Insightinto processesmodifying interstellar hydrogen can be em• glow, but they should be used with caution when providedby Lyman alpha measurementsfrom the Voyager appli• to neutral atoms •side the •ah's orbit. Two andPioneer spacecraft at largedistances from the Sun [e.g., additiomlio•zation processesmay b•ome im••t • the Hall et al., 1993] and by proposeddirect measurementsof Sun's vic•ty • < 0.5 AU): io•zation by solar w•d interstellarhydrogen atoms [Gruntman, 1993] and hello- el•trons [e.g., Ruci•ki at• Fahr, 19891 •1 two-step sphericENAs [Gruntmatt,1992]. photoio•zationby solarphotons [Gruntman, 1990]. These The abovearguments do not meanthat the hot model has two eff•ts c• be disregard• for the pu•se of the present completelylost its validity and usefifinessfor hydrogen. It work. is of little help in detemfininginterstellar hydrogen charac- Accord•g to the "hot" model, numberdensity distribu- teristicsin the unperturbedLISM, since a detailedmodel tions of •terstellar atoms• the heliosl>herecm• be given by allowingfor the effectsof the heliosphericinterface has yet to be developed. However, it can be reliably usedfor the n.,.c(R,O) = n.,.c(oo)f.,,•(R, O) (3) kind of estimatesrequired in this work providingcertain modifications are marie. The straightforwardway to wherethe numberdensities are scaledby their corresponding improvethe modelis to usedifferent values of velocitiesand unperturbedISG values, n.,.,(oo). Functionsf.,.•(R,O) temperaturesfor the LISM hydrogenand helium components have been calculatedby many authors,[e.g., Fahr, 1974; at inf'mity. For helium the parmnetersof the unperturbed Meier, 1977; Wu a•wlJudge, 1979]; the computercode used ISG can be used, and for hydrogenthe effectivevelocity and in this work was developedby D.Hall (private conmm- temperatureconsistent with the interplanetaryglow experi- nication, 1992). ments(performed in the inner solar system)should provide A selectionof the ISG paran•etersat infmity, number reasonableresults. We also use a higher value of the density,velocity, and temperature, requires special consider- helhnn-to-hydrogennmnber densityratio than existsin the ation. Until recently,it was a conunonassumption that the unperturbedLISM, the latter is believed to be close to GRUNTMAN: NEUTRAL SOLAR WIND - GEOMAGNETIC STORM WARNING 19,217

-0.1. Such modifications shonld allow one to continne to 1.0E-02•! use the well developedand relativelyshnple hot modelof • 1.0E-03• the ISG for our purposes. 'E An apparentgap exists between modelling the transfomm- tion of interstellarhydrogen characteristics in the unper- o1.0E.0•I turbedstate in the LISM and the stateof hydrogenstill far 180' away from the Snn, bnt after its passagetlu'ough the 1.0E-07,1 terminationshock. One shehidkeep in nfindthat currently even the mostdeveloped model of Barahoyatut Malama 1.0E-081 [1993] is limited since it does not take into consideration J••1.0E-09][ • 1.0E-10 magneticfield, cosmicrays, and temporalvariations (helio- e- sphere"breathing"), and it is restrictedto the caseof a two- shockconfiguration. Incidentally, this gap in modelinghas an importantimplication for heliosphericstudies. Most of 1.0E+00: the experimentaldata on interplanetaryglow are obtained from spacecraftnear the Sun (R < 5 AU). Consequently, all attemptsto improvethe best fit of experhnentaldata to 180 ø the unperturbedISG parametersby an elaborationof the hot 'O• 1.0E-01 model (for example by introductionof anisotropyand temporalvariation of the solarwind and solar Lynmnalpha 120o radiation) have limited value. The velocity distribntion functionof hydrogenatoms far away from the Sunbut inside ß- 1.0E-02 the terminationshock is essentiallynon-Maxwellian, being determinedby the structureand physicalprocesses in the E heliosphericinterface. lb) In our calculationsthe followingparan•eters are nsodfor 1.0E-03 interstellarhydrogen and helium atoms: 0.0 0:1 0:2 0:3 0:4 0:,5 0:$ 0:7 0:8 0:9 1.0 dislance from Sun. AU n.(oo) = 0.1 cm'3 nm(oo)= 0.02 cm'3 (4a) VH(oo) = 20 km/s VH•(oo)= 26 hn/s (4b) Figure 2. Radialdependence (0 < R < 1 AU) of the numberdensity of interstellar(a) hydrogenand (b) helimn T.(oo) = 14,000 K T.•(oo) = 8000 K (4c) for 0 = 0, 60ø, 120ø, and 180ø. Read 1.0e-02 as 1.0 x 10'2. •. = 0.8 •.• = 0 (,kl)

The ionizationrate of heliumatoms, which is ½letem•edby exchangecross sections and higherhydrogen atom munber photoionization,is f/.•(R•) = 0.8 X 10'7 s'l at the Earth's densities. orbit. The photoionizationrate of hydrogenatoms is f/•H(R•) = 2.0 X 10'7 s'l, the chargeexchange ionization 3.2. Outgassingof Interplanetary Dust rateis f/•X(Rz)= t•ssw,H+q.+,. = /•.+ Vswq.+,.= 3.75 Interplanetaryspace is filled with a populationof dust x 10'7 s'l, and the total lossrate /•H(Re)= /•x + /•HP.= whichtends to congregatetoward the Sun and is largely 5.75 X 10'7 s'l. Solarwind parametersat 1 AU are as- restrictedto within 15ø of the eclipticplane [Ba•s, 1971; sumedto be N•H+ = 5 cm'3 and Vsw= 500kan/s; the Holzer, 1977; Fahr et al., 1981]. The surfacelayer ( • chargeexchange cross section is qH+,H= 15 X 10'• cm2. 500,• ) of thedust grains is quickly saturated bybombard- The calculatedradial dependentnumber densities of inter- ing solar wind ions followed by desorptionof neutral stellar atomsare shownin Figtires2a and 2b for • = 0, particles,atoms and molecules,from the surfacein order to 60 ø, 120ø, and 180ø. Differences between distributions of maintainequilibrium. The saturationtime doesnot exceed hydrogenand heliumare due to the effectsof solarradiation few years which is much less than the corresponding presstireand ionizationloss. One can see that interstellar Poynting-Robertsonresidence time [e.g., Fahr et al., 1981]. hydrogen is strongly ionized and only a small fraction Banks[1971 ] consideredthis process in detailand calculated reachesthe region insidethe Earth's orbit. The downwind numberdensities of hydrogenand heliumatoms due to dust region(around 0 = 180ø) is practicallydevoid of interstel- outgassing.Neutral particle number density is sealedby an lar hydrogenatoms. For interstellarhelium the situationis interplanetarydust geometricalfactor F(R), which is the reversed. Helium numberdensity is substantiallyenhanced inverseof the solarwind ion free pathagainst interception in the downwindregio n by gravitationalfocusing, and the by interplanetarydust grains. The uncertaintyof this smallerionization rate allows helium atomsto easily reach parameter, as discussedby Banks [1971], is 4 orders of the inner part of the Earth's orbit. magnitudeF(1 AU) = l0 '21-- l0 '17em 'l. Holzer [1977] The numberdensity of interstellarhelium atoms is higher favoredthe value F(1 AU) = 2 X l0 '19em '• which was thanthat of hydrogenatoms although the helium abundance later adoptedby Fahr et al. [1981]. is much smaller than that of interstellarhydrogen in the While helium desorbsfrom grains in atonficform, the unperturbedLISM. The relativeincrease of the interstellar physicsand chenfistryof hydrogendesorption are poorly helium abundanceinside Earth's orbit requires one to understood.Banks [1971] initiallyassumed that dust emitted considereffects involving collisions of solar wind ions with hydrogenis in atomic form. Later Fahr et al. [1981] neutralhelium atoms. Helium atomscan oftenbe disregard- reconsideredhydrogen desorption and conchidedthat most ed further away from the Sun due to nmch smallercharge of hydrogenis desorbedin molecularform. An hnproved 19,218 GRUNTMAN: NEUTRAL SOLAR WIND - GEOMAGNETIC STORM WARNING model of dust distribution was also used in the latter ionizationrates, as those adopted by Siscoeatul Mukherjee calculations,and we will asstimehere (with somemodifica- [1972],the total molecular hydrogen ionization rate is 2.85 tions introducedbelow) dust-generatednumber density x 10'7 s '! at 1 AU. Thetotal molecular hydrogen loss rate, distributionsof hydrogenand helium as givenby Fahr et dueto photodissociationandionization, is 5.95 x 10'7 s '!, al. [1981]. ascompared to 6.6 x 10'7 s '! usedby Fahret al. [ 1981]. The heliumnumber density, as calculatedby Fahr et al. The probabilityto' of the photodissociationbranch is now [1981] (their Figure 5) and adoptedas basisin tiffs work, to'= 0.52 ascompared tothe tc= 0.05given by Fahret can be approximatedas al. [1981]. Thesenew rates are used here to modifythe hydrogen -- 2.32 x 10'5 x R'• cln'3 (5) numberdensities of Fahret al. [1981] in thefollowing 0.0232 AU < R < 0.232 AU manner. If we assumethe productionrate of molecular hydrogenof Fahret al. [1981], thenthe H2 mlmber density rtl_l½(R) -- 10-4cm-3 (6) wouldbe a factor(ratio of lossrates) 6.6/5.95 = 1.11 0.232 AU < R < 0.464 AU higherthan the numberdensity obtained by Fahret al. [1981]as can be derived from (9). Asstintingthe stone loss n.c(R)= 2.77 x 10-5 x R'!'67 cm '3 (7) rate of atonfichydrogen, the atonfichydrogen number 0.464AU < R < 1 AU densitywould be a factor tc'/tc= 10.4 lfigherthan the one givenby Fahret al. [1981 ]. Suchcorrections areobviously where R is in A U. only approximate(within 10-15 % accuracywlfich is The hydrogennumber density can be approxhnatedas acceptablefor the purposeof the presentwork) since accuratecalculations require a detailedand complicated n.(R) = 4.13 x 10'6 x R-l'15 cnf 3 (8) modelas usedby Fahr et al. [ 1981]. The correcteddust-generated number densities of atonfic for 0.0232AU < R < I AU, where R is in AU. The andmolecular hydrogen, assumed in thiswork, are distributionof molecularhydrogen can be easilyderived from that of atonfichydrogen. The lossof H2 occurs n.(R) = 4.3 x 10'5 x R'•'!5 cm '3 (12) mostlyin twoprocesses: photodissociation with probability to, and photoionizationwith the probability (l-t0. The llH2(R)-- 3.9 x 10'5 x R'!'!5 cm -• (13) former process, H2---> H + H, (9) where R is in AU. The radialdependence of dust- producedneutrals, n., rtH2, and n.c, is givenin Figtires3a- is asstunedto be the onlysource of atonfichydrogen. In 3d wherethey can be comparedwith the interstellar neutral equilibriumthe production rate of hydrogenatoms is equal atomdistribution at differentangles O. Obviously,the to their loss rate, dust-producedneutral particle distributions are independent of the angle 0 andare identicalon thesefour figtires n"2 x /•"2 x 2 x K = n. x /5. (10) (Figures3a-3d). One can see that interstellar helium is much more abun- and consequently, dantthan dust-produced helium, and the latter can be safely neglected.Dust-produced hydrogen atoms are more numer- n"2 = n.x /5.(/5.2 x 2 x K), (11) otisin the Sun's vicinity while interstellar hydrogen becomes muchmore abundant at 1 AU fromthe Sun. Thepoint at where n.2 and ]•H2 are molecularhydrogen number whichcontributions from thesetwo sourcesare equal densityand loss rate, respectively. dependson the angle 0 andvaries from 0.35 AU for 0 The hydrogennumber density distribution (8) can be = 0 to 0.8 AU for 0 = 180ø. Dust-producedmolecular easilymodified on the basisof tipdatedrates and cross hydrogenis concentratedin the regionclose to the Sun. sections.Fahr et al. [1981]used the molecularhydrogen Incidentally,its presence should be manifested bya molecu- photodissociationrate (at 1 AU), 0.34 X 10'7 s'z, and total lar hydrogencomponent in solar wind piclalp ions. Mea- ionizationrate, /5.2 = 6.3 x 10'7 s'!; the latter rate was surementof molecularhydrogen piclmp ions should allow suggestedby Siscoeamt Mukherjee[1972]. Theserates theindirect determination of the amount of interplanetary determinedthe photodissociation branch probability, t• = dustin the Sun'svicinity. 0.05. No directexperimental data of dust-producedneutrals are The photodissociationrate, however,should be higher available,and the tincertaintyof their numberdensity since,as pointedout by Dalgarnoaml Allison[1969], estimatesremains high. Opticaldetection of dust-generated contributionsfrom photon absorption into discreet states that neutralparticles is difficultbecause the light scattering predissociatewere apparentlyignored. D.E.Shemanskyregion is veryclose (only 5ø-10 ø offsetangle) to the Sun. (unpublishedmanuscript, 1991) thorouglfly investigated these The experhnentalconcept to optically(584 and 1216•) additionalbranches and reconunendexl thephotodissociation measure such neutrals was developed by Fahret al. [1980]. rate, 3.1 x 10'7 s '! at 1 AU [Shemanskyamt Hall, 1992]. Twosounding rocket (Skylark 12) launches were attempted For assumedsolar wind parmnetersand ulxlatedcross in 1985and 1989from the range"Barreira do Inferno"in sections[Barnett et al., 1990], the H2 ionizationrates at 1 Natal,Brazil. Thefirst launch failed due to theproblems AU are 0.001 x 10'7 s'• and 1.2 x 10'7 s'! for solar wind withthe separation of the third rocket stage. Experhnental protonimpact and charge exchange correspondingly. data obtainedduring the second launch in 1989 were Assumingthe same H 2 photoionizationand electron hnpact inconclusivebecause of a malfunctioningon-board computer GRUNTMAN: NEUTRAL SOLAR WIND - GEOMAGNETIC STORM WARNING 19,219

andunexpectedly strong signal contanfination by ltigh-energy particles [Lay et al., 1989, 1991]. The feasibility of a 1.0E+00• direct, in sire measurementof dust-producedneutral parti- cles on the flittire Solar Probe ntission is tinclear due to spacecraftprotective shield and ablation[Grtozt- O J:::1.0E.0• t rH H man, 1993]. 3.3. ENAs From the Heliospheric Interface The solar wind is a higldy supersonicplasma flow into the LISM, which is characterizedby a certainfinite pres- sure. It is believedthat this supersonicplasma flow temti- 'j•'01='08.0E-07 natesat a solarwind ternfinationshock front, beyondwhich its kineticenergy is largelyconverted into thermalenergy in 1.0E-08 [a) subsonicplasma. As neutral ISG atoms passthrough the interfaceregion, there is a certainprobability for solarwind protonsto charge exchangewith interstellargas and give 1.0E-01 •..•. rise to heliosphericENAs (HELENAs). This processwas first invoked by Patterson et al. [1963]. A detailed • •.oE-o2•• HELENA fitix sinrelation was performed recently by 1;::1.0E-03• H Gruntman[1992]. The expectedHELENA flux at 1 AU is highlyanisotropic with an intensityof abont 200 cm':s '• sr'• at its maximum in the upwind direction, 0 = 0. The (I)1'øE'05 t kineticenergy of the hydrogenatoms is in the range 200- dust 1000eV. The upperlintit of the HELENA numberdensity within1 AU is 10's cm '3. Thisvalue is comparablewith the dust contribution and much smaller than the ISG contribution (b) for R > 0.5 AU. For R < 0.5 AU the HELENA contributionis muchsmaller than that of interplanetarydust. •.OE-O1ISG - He Consequently,the contribution of HELENAscan be neglect- ed in this work. H '?''OE'02 t "'"- H 2 o 1.oE.o31 4. Stationary Neutral Solar Wind Two differentprocesses contributing to the productionof 1.OE-04• a neutralcomponent in the solar wind are recombinationin the expandingplasma, and chargeexchange of the solar 1.0E-O• wind ions on neutralparticles in interplanetaryspace. Not dust all NSW atoms would reach the Earth because of the .OE-O7t ionizationloss on their way from the placeof their "birth" c- 1.oE-o8 (c} to 1 AU.

4.1. Probability to Reach 1 AU 1.0E-01 When a neutral atom is born in the solar wind it is "? •.OE-e2 E movingin outwarddirection with the local velocityof the o 1.0E-03 solarwind. An atom is a subjectto lossprocesses, and it may not stirviveits travelto the Earth. The probabilityfor - 1.0E-04 an atom born at a distance R0 from the Sun to reach J•1.0E-0•:, Earth's orbit RE is HG R• dust E P(Ro,V,0 = exp .... dR (14) :::1 1.OE-O8 (d) v•(R) 1.0E-Og Ro 0.0 0.1 0.2 0:3 0:4 0:5 0:5 0:7 0:80:g 1.0 distance from the Sun, AU whereionization rate /3• andatom velocity V• dependon thedistance from the Sun, R. Assunfingthat atom velocity is constantand /3• = /3a(R•)(Rœ/R) 2 , one can transform Figure 3. Numberdensity radial dependence (0 < R < 1 AU) (14) as of dust-produced("dust") helium and atonticand molecular hydrogen. Dust-producedneutral particle distributions are independentof the angle O. Interstellarhydrogen (ISG-H) P,•(Ro,Va) = exp [_ 13'•(Rtr)R•r(------Rtr -- 1 ) (15) andhelium (ISG-He) numberdensities are shownfor angles v• R0 (a) 0 = O; (b) 0 = 60ø; (c) 0 = 120ø; (d) 0 = 180ø . The ionizationrate of atonfichydrogen that is usedin the tlS) is deternfinedby photoionizationonly. Chargeex- 19,220 GRUNTMAN: NEUTRAL SOLAR WIND - GEOMAGNETIC STORM WARNING

change,which is the main ionizationloss process for atonfi½ whereN•(R) = 1•.+ (Re/R):is thelocal solar wind electron hydrogenha haterplanetary space, does not contributeto the numberdensity. For our purposes,the solar wind can be lossof the N SW atoms. When chargeexchange octOrs, it assumedto be fifily ionizedplasma with equalelectron and resultsha a replacementof one neutralatom moving with the protonnumber densities. The recombinationrate coefficient solar wind velocity by anotherone moving with the same ot dependson the electrontemperature T• , which in turn velocity and in the same direction. Some momentran is a fi•nctionof the distancefrom the Sun. Only hydrogen exchangetlmt may happenha such a collisioncould contrib- atomsare born in the recombinationprocess. The recombi- ute to the angularbroadening of the NSW flux. nation-producedhydrogen ENA fitix at Re is equalto an The effect of scatterhagin chargeexchange collisions can integral, over the distancefroln the Sun to Earth, of the be easilyestimated if one notesthat the scatterhagangle (for productionrate multiplied by the probability P. to reachRe small angle elastic scatterhag)depends on the collision without ionization and redaicedby the factor (R/Re)2 impact parameter b approxhnatelyas = 2 U(b)/E•, allowing for the divergenceof the NSW tlux. The station- where U(b) is the hateractionenergy of collisionpartners ary NSW flux (cm'2 S4) at Re is at the distance b and Ee is the projectileenergy. For resonancecharge exchange(H + + H--> H + H + and RE He ++ + He- > He + He++), typical hateractionenergy is few eV (impactparameters 1 - 2.5 3,)restilting ha scatter- ing anglesof few milliradian (few tenthsof a degree). The •NSW,reeom--'(N•H +)2]ol(te) (RE/R)2 PH(R' Vsw) dR(16) scatteringis much more importantfor proton charge ex- Rmin changeon helium atoms(H + + He -- > H + He+) when typicalinteraction energy is 10-15 eV resultingin scatterhag where the electrontemperattire depends on the distancefrmn angles 1ø-2ø. The scatteringeffect is disregardedha this the Sun, T• = T•(R). work. We will use a typical radial dependenceof the electron The probabilitiesPH,m for NSW hydrogenand heliuln temperatureas given by Holzer [1977] Offs Figure lb), atomsto reach 1 AU withoutloss are shownha Figure 4 as which can be approxinmtedby a fi•nction of the distance R0 of the place of their birth from the Sun; the solarwind velocityis Vsw= 500 lmds. T = 1.32 x 10•K (17) Theradial dependence is strong at theSun's vicinity wlfile R < 0.056 AU the probabilitieschange only slightly at larger R0. For atomsborn at R0 > 0.1 AU the probabilityto reach1 AU T = 1.23 x l0 s x R'ø-•2 K (18) is higherthan 80% for heliumatoms and 55 % for hydrogen 0.056AU < R < 1 AU atoms. The higher atom velocity VA the smaller is ENA ionizationloss. The solar wind electronhnpact ionization where R is in AU. The hydrogenradiative recombination [Rucinski amt Fahr, 1989] would contribute to the NSW rate coefficientha the temperaturerange of haterestcan be atom lossrate haa way sinfilarto the ionizationof interstel- approximatedby lar anddust-generated neutrals. The situationis differentfor two-stepphotoionization [Gruntnmn, 1990]. As a restlitof 7.9 X 104 X T•-•'2 cmss '• (19) their high-velocity, NSW atoms are Doppler-shiftedand l0 s < T• < 3 x 106K two-stepphotoionization is negligible. 4.2. Recombination 7.9 X 10'•ø X Tffø's cmss '• (2O) 10n < T, < 10SK The number of neutral hydrogen atoms born ha the recombinationprocess ha a unit volumeduring time interval where T, is ha Kelvin. For solar wind parmnetersand dt at the distanceR from the Sun is a(T3 •(R) dr, ionizationrates given above, the hydrogenENA fitix in the solarwind due to recombinationwould be F•sw,•½om= 54 cm'2s'• at theEarth's orbit. In termsof a nulnberdensity fractionthe recombinationproduced neutrals constitute 2 x 10'7 of the quiescentsolar wind at I AU from the Sun. 0.8 Q- 0.7 4.3. Charge Exchange ._• 0.6 The two most importantsolar wind ion componentsare • 0.5 protonsand doublycharged helimn ions (alphaparticles). Major componentsof theneutral gas participating ha charge -QQ. 0.41O.3 exchangecollisions are hydrogenatoms, hydrogen mole- 0.2 cules,and heliumatoms. The chargeexchange processes 0.1 contributingto the N SW productionare 0..0 0:1 0:2 0:3 0:4 0:5 0:6 0:7 0:8 0:9 1.0 H + + H -- > H + H + (21) distance from the Sun, ^U H + + H2 --> H + ... (22) Figur• 4. ?robabilityfor neutral hydrogenand helium H + + He --> H + He + (23) atomsto reach1 AU as a fi•nctionof the atoln'sbiffi•place; He ++ + H 2 -- > He + 2H + (24) atom velocityis 500 •ds. He ++ + He --> He + He ++ (25) GRUNTMAN: NEUTRAL SOLAR WIND - GEOMAGNETIC STORM WARNING 19,221

The crosssections for processes(21)-(25) are well known NSW flux. The totalflux (cm'2 s'•) of hydrogenand helium [e.g., Banu•tt et el., 1990], and they are shownin Figure 5 atomsat 1 AU fromthe Sun, F•sw..,.• , is anintegral over for collisionvelocity range from 300 to 1400 kan/s. Several the distance from the Sun to Earth branchesare possiblein the processes(22) and the total crosssection for charge transfer to a proton in a collision FENsw,H with a hydrogenmolecule is given here. For a collision velocity of 500 knffs the chargeexchange cross sections are =NC•. +Vsw [P.(R, Vsw) En•(R) q.+,•(V•v) dR (27) Om+,H = 15.8 X 10'•6 cm 2 groin qH+,H2 = 4.9 X 10'•6 cm 2 q.+,Hc = 0.01 X 10'•* cm2 E •qSW,He q.c++,.2 = 0.27 X 10'• cm2 q.c++,H• = 2.4 X 10'• cm2 (28) for processes(21)-(25) correspondingly. k We assumethat the solar wind sphericallyexpands with =N•.•+ +Vsw [P.c(R, Vsw) En•(R)q. .... •(Vsw) dR the constantvelocity Vsw; its neutral componentis much smallerthan the ion component,and the efl•.ct of the Sun's gravitationalfield is negligible. Let us considera solarwind The lowerintegration linfit usexlby us is ginin -- 0.05 AU. elementat the distance, R (0 < R < RE), from the Sun. Actually,the resultsof calculationsare only slightlysensi- The probability dpt for a selectedion I (H + or He ++), to tiveto thevalue of Rmin sincethe probability, P.,.•(R, V), chargeexchange while moving from R to (R+dR) is decreasesrapidly with the approachto the Sun(Figure 4). In caseof theNSW hydrogencmnponent, F•sw,. , charge dp. = dR x E n•(R)q,k(Vsw) (26) exchangeoccurs on interplanetaryH, H2, and He, wlfile in k caseof the NSW heliumcomponent, F•sw,.• , neutral particlesinvolvexl are H2 and He only. The alpha where nk are the nmnberdensities of neutralspecies and particle componentof the solar wind is assumexlto be qt,k are correspondingcharge exchange cross sections, and • = N•.•++/N•+= 0.05and abundance of singly charged the summationis performedover all presentneutral particle heliumions He+ is negligiblysmall. species,k (H, H2,and He). It is alsoassumed here thai the Theangular dependence of stationaryNSW hydrogenand velocity of neutral atoms is much smaller than that of the heliumcmnponents, F•sw,.(O ) and F•sw,.•(O),is shown solar wind, and hencethe collisionvelocity between parti- in Figure 6. Also shown are contributionsto the NSW cles is equalto Vsw. A contributionof the chargeexchange hydrogencomponent due to chargeexchange on interstellar betweenR and (R+dR) to the N SW flux (cm'2 s'l) at the hydrogen(ISG - H) andhelium (ISG - He) atomsas well as Earth's orbit wouldbe a productof the local solarwind ion theirsum F•sw•.(ISG).For parametersaccepted in tiffs flux, Fsws(R), probabilityof chargeexchange, dpt, workmost of thecontribution to theN SW hydrogencompo- probability for neutral atoms to reach the Earth's orbit nent is due to charge exchangeon interstellarneutrals: without ionization, P.,.•(R,Vsw), and factor (R/RE)2 chargeexchange on interstellarhydrogen dominates every- allowing for the divergenceof the sphericallyexpanding whereexcept the downwindregion where charge exchange on interstellarhelium becomes more hnportant. The NSW hydrogencomponent due to clmrgeexchange 1.0E+02• on dust generatedneutrals does not dependon angle t9, constitutes 9.2 x 102 cm'2 s'• and its relative contribution becomessubstantial in the downwindregion. Most of the 1.0E+01 2 chargeexchange occurs on dust-generatedetontic hydrogen (approximately three fourths) and some on molecular

1.0E+00 hydrogen(one fourth) with negligiblecontribution of dust- generatedhelium. The NSW helium component •sw,ttc is generatedalmost exclusively by chargeexchange on inter- 1.0E-01 stellarhelium atoms. A contributionof chargeexchange on dust-generatedneutrals is lessthen 4 cm'2 s'• and can be disregarded. 1.0E-02 Most of the expectedstationary NSW hydrogenatom flux, with the exceptionof the downwindregion, is pro- 1.0E-03 ducedby the interactionof the solarwind ionswith interstel- 0 400 800 1200 1600 lar neutrals. One has to note, however, that characteristics velocity. km/s of interplanetarydust in the Sun'svicinity are not known Figure 5. Velocity dependenceof the charge exchange with highaccuracy. If thedust population is morenumerous crosssection(10 '•6cm2). (1) H + + H--> H + H + ; than adoptexlin this work, then the situationmay change, (2) H + + H 2 --> H + ...; (3) H + + He- > H + He+; and the dustcontribution could become the mosthnportant (4) He ++ + H2 --> He + 2H + ; (5)He ++ + He--> contributingfactor detemfining the properties of theNSW at He + He + + 1 A U from the Sun. 19,222 GRUNTMAN: NEUTRAL SOLAR WIND - GEOMAGNETIC STORM WARNING

1E+04 laG031.As a CME movestoward the Earth, its radialsize F ENSW, H grows,it decelerates,and different parts of the CME move with differentvelocities. The CME leadingedge reaches the Earth (at the time momentt•) and thenafter sometime (at Ho 1E+03 the time momentt2) the trailingedge passes the Earth. At this moment, t2, the trailingedge is at P4 = 1 AU, the leadingedge is at P3, and the CME radial size is • = P3- P4 (Figure 7). An advantageof such a model, which significantlysimplifies the computation,is that at any given 1E+02 Fi•SW,H(ISG) momentof time ionsat a givenpoint in the CME haveonly one radial velocity and converselythe ions with a given velocitycan be foundonly at one pointof the CME. ISG - He ISG - H We will assumehere that p• = 0.1 AU, P2 '- 0.05 AU, 1E+01 0 30 60 90 120 150 180 P3 = 1.1 AU, and P4 = 1.0 AU. The radial size of the angle theta, degree CME changesfrom A, = p•- P2 = 0.05 AU to A2 = P3- P4 = 0.1 AU. The initial velocityis assumedto be V0 = Figure 6. Angular dependenceof the stationaryneutral 1200 km/s and the decelerationof the leadingedge, a• = solar wind (NSW) hydrogenand helium flux (cm'2s'•) a(p0 = - 2.3 x 10'3 km s'2, thevah•e adopted by Hsiehet componentsat the Earth's orbit; Fn•sw,.and F•,sw,.care al. [1992]. The propagationtime requiredfor the leading thetotal fluxes; F•sw,.(ISG) is the total contribution of the edgeto reach P3 = 1.1 AU and the trailing edgeto reach ISG consistingof chargeexchange on interstellarhydrogen 1 A U, t2 - to = t2 , can be obtainedfrom (ISG-H) and helh•m(ISG-He). a, t222/ 2 + Vot2 - ( p3- p, ) = 0

5. ENAs due to CME whichgives t2 = 1.45 X 10• s. The accelerationof the Many typicalfeatures of CMEs were describedrecently trailing edgeelement can thenbe obtainedfrom in detailby Hildner [1992], Gosling[1992, 1993a,b], and Hundhausen[1993]. For otlr modelingit is importantthat ae = a(p3 = 2 [(304-P3 - Vot2I / 4 = - 3.01 kax#s2 (30) a largenonrecurrent geomagnetic storn• can be prrxlucedby a strongCME with the leadingedge velocity > 1000krafts In our model the accelerationof any given CME element near the Sun. CMEs often spreadmore than 45ø in dependslinearly on the initial positionof thiselement, p heliocentriclongitude and latitude,and they are character- < p < p•), that is, ized by increasedabundance of helium (He+ +/H +) and a(p) = o.2+ (p - 02) (a, - a2) / (p, - P2) (31) depressionof plasma ion and electron temperatures. Considerationof the effectsof shockspreceding CMEs as Time t2 correspondsto the mmnentwhen the CME trailing well as plasmabetween the shocksand CMEs are beyond edgereaches the Earth. The magneticstorm starts earlier at the scopeof the presentwork. We assrunehere that a storm time, t, , when the leaclingedge reachesthe Earth. This startswhen CME plasmahnpinges on the magnetosphere. time moment t I Canbe found from The model of CME propagationchosen here uses an approachclose to thatof Hsiehet al. [1992], the latterbased a, • / 2 + Votl - (104'Pl) -- 0 (32) onrestilts of thecomputer sinrelation by Hanet al. [1988]. Several substantial modifications are, however, introduced whichgives t, = 1.28 x l0 s s. CME/ENAsborn at Pl by us to facilitatethe computation. at the moment to would be the first to reach the Earth at It is assumedthat at the moment of thne, to = 0, all the moment, t3 , and are the earliestpossible CME/ENA partsof the CME has the samevelocity V0, the CME's precursorof the approachingCME. Tlfis momentis in our leadingedge is located at Pl, its trailing edgeis at P2, case and its size in a radial directionis A. = Pl - P2(Figtire 7). Each CME element moves with a certain constant deceler- t3 = (P3-P,)/ Vo = 1.12 x lOss (33) ation, a(p) < 0 , which is a linear fimctionof the initial positionof tlfis element, p(to). The CME leaclingedge is The maximumpossible time delay betweenthe CME/ENA movingfaster than the trailingedge, that is laCo,)l < advancewarning and the onsetof the stormis consequently t•- t3 = 1.6 X 10•s • 4.4 hours. t=t 0 t=t 2 For assumedinitial CME velocity, V0 = 1200 kroy's,the CME leadingedge hnpinges on the Earth with the velocity, V, = V0 + a• t, = 905 km/s, and the trailing edgehas the AI I ...... a velocity, V2 = V0 + th t2 = 763 knffs at I AU. The calculationof the probabilityfor a hydrogenCME/- ENA to reach 1 AU withoutionization is more complicated P2 Pl 04=1 AU P3 than in case of the stationaryNSW hydrogenatoms. A CME/ENA is a subjectof photoionizationin the sameway Figure 7. A model for coronalmass ejection (CME) as NSW hydrogen. However, a contributionof the charge propagationtoward the Earth. exchangecannot be neglectedanymore. An ENA moves GRUNTMAN: NEUTRAL SOLAR WIND - GEOMAGNETIC STORM WARNING 19,223

relativethe graduallydecelerating CME plasnmand at some From a computationalpoint of view it is convenientto moment of time crosses the CME's leading edge and evaluateintegrals (37) and (38) by selectingan elementin continuesits flight in the stationaryunperturbed solar wiml. the CME at to = 0 and consideringthe contributionof this We assumehere that no ionizationdue to chargeexchange elementto the NSW flux during its propagationtoward 1 occurswhen an ENA is insidethe CME. This assumption AU. The sum of contributionsof all elementsinitially is reasonable,since if the chargeexchange does occur, a constitutingthe CME wouldgive the CME/ENA flux. The new ENA would have velocityclose to that of the original neutral particle radial number density distributionsare ENA. Outsidethe CME, an ENA is a subjectof charge known. The numberdensity radial dependence of the CME exchangeas a neutralatom moving relative the solarwind ions and electronsis different from that of the stationary plasmawith the velocity, VA- Vsw, where VA is theatom solar wind. At 1 AU the CME number density is not velocityrelative the Sun. particularlyhigher than that of the stationarysolar wind If an ENA is born with the velocity VA at the distance [e.g., Gosling, 1992]. To keep the model simple, we R• from the Sunand crosses the CME's leadingedge at the assumehere that the CME's nt•mber density is Na•+= 10 distance,R 2, thenthe probabilityto reachthe Earth(at Rs) cm'3 at 1 AU anda solidangle subtended by the CME does without ionization is not changewith the propagationfrom the Sun. Such an assumptionleads to a simple relation between plasnm P(RI ,R2,VA) = PpH(Ri,VA) X Pcx(R2,VA) (34) numberdensity at 1 AU and numberdensity at arbitrary R, where takinginto accountthe radial expansionof the CME's size, A. The CME number density would be approxilnately P•.(R , , V3 factor 2 larger at 0.1 AU thanit wouM be for a stationary expansioncase. In reality, one can expect higher enhancementof the = exp ß . ,. (1 ------) (35) CME plasmadensity closer to the Sun sinceCMEs experi- ence greater expansionin transit to the Earth than the stationarysolar wind [e.g., Gosling, 1992]. The correction allowing for this effect will result in higher CME/ENA (36) fluxes, especiallyfor the early arriving neutral particles, than obtained from the present model. Therefore our = exp , , , (1- estimates of the CME/ENA flux can be considered as [N•.+ (Vz-VzVsw) q.+,aRe RsR2 ] conservative. where q.+,. = q.+,.(Va - Vsw). Thedistance R: canbe The alphaparticle component in a CME is assumedto be •sily found from conside•g the ENA motionwith the • = Nsm++/NSa+= 0.10, a factor 2 higherthan in the comet wloci• Va •d the d•elerat•g motionof the quiescentsolar wind. There are indicationsthat the ratio CME's l•d•g •ge. He +/He+ + may increasein CMEs up to suchhigh valuesas As • caseof the smtion• NSW, thereare two processes > 0.1 [e.g., $chwennet al., 1980]. The chargeexchange that cont•bute to the CME/ENA flux: r•omb•ation •d crosssections of helium ions and atomsare approximately c•ge exch•ge on •te•l•eta• neutrals.•m CME/ENA factor 3 higher for singly chargedions than for doubly flux will be characteriz• by a ch•g•g • tinm wlocity chargedions in the energyrange of interest. Thereforethe dist•bution •nction. •e flux due to clmrg• •xchang•, contributionof singly chargedhelium ions to the helium G•H• (cm'2 s'• ), with velocity V at 1 AU at a thne CME/ENA flux may constituteup to 30 % of the total moment t is helium ENA flux. We further disregardthis contribution. The electrontemperature is known to be depressedin CMEs, andwe adopthere the temperatureradial dependence (37) T• = 5 X 10• R'•K (39) Re-R where R is in AU. The recombination rate coefficient 0• :[ N•R, V,t-v k is approximatedby (19) and (20). R.•in As an example, the calculatedtinhe dependence of the CME/ENA fluxes, Ga and G.., is shownin Figures8a where Nt[R,V,t-(Re - R)/V] is the radial and velocity and8b for 0 = 0ø and 180ø, correspondingly.The CME distribution of the CME's ions at the thne moment plasmaimpinges on the Earth (i.e., the geomagneticstorm It - (Rs- R)/V]. The CME/ENA flux due to recombination starts)at the time moment t = 62 (one time unit, or time would consistof atomic hydrogencomponent only and is step, is 300 s = 5 rain; t = 0 doesnot coincidewith the equalto moment to). The CME's trailing edgepasses the Earth at t = 120. The CME/ENA hydrogenflux (Figures8a and (38) 8b) consistsof contributionsof recombination("recom") and chargeexchange on interstellar("isg") and dust-generated 2 ("dust") neutrals. a(r3 P½,V,O(R/RD' ,m We are interestedmostly in ENA flux precedingthe beginningof the storm (t < 62). One can seethat the dust- generatedneutrals account for themajor part of thehydro- 19,224 GRUNTMAN: NEUTRAL SOLAR WIND - GEOMAGNETIC STORM WARNING

1.6E+04 Figtires9a and 9b reflect the differencein relativecontribu- 1.4E+04. storm starts tions of chargeexchange on interplanetaryhelium and (a) hydrogenatoms. The energydistributions of heliumatoms A 1.2E+04- aresimilar to thoseof hydrogenand are shown in Figure10. Until the CME impingeson the Earth, the CME/ENA • 1.0E+04- flux arrivessimultaneously with the quiescentNSW ENAs. A 8.0E+03- Althoughthe CME/ENA flux is slightlysmaller than that of thequiescent NSW, energies of CME/ENAsare significantly higher.This energy dependence on ENA originallows one o6.0E+O3.reco m to easily distinguishbetween the NSW and CME ENA •- 4.0E+03 fluxes. The restiltsof the CME/ENA shnulationare model sensi- tive. Theyshow, however, the majorcharacteristic CME/- o.0E + 00 / ENA features(Figtires 8, 9, and 10) as well asdemonstrate o 20 4• 60 80 100 120 14o the relative in•portanceof the processesinvolved. For time, 300 s higherCME numberdensity and lower electron temperature 2.5E+04 the relativecontribution of recombinationwould be larger, stormstarts althoughit is unlikely that it maybecome the dominating 2.0E+04 (b) process.The major source of uncertaintyin esthnates • • remainsinterplanetary dust.For example ifdust geometrical <• factor, I'(R)' is higherthan adopted in thisiDaper, then it • 1.5E+04. willrestlit in higherENA flux preceding the storm, the latter • beingin,portent for storm'sadvance warning.

E o 1.0E+O4- x• 1.2E+04 • 5.0E+03- /t= 60 50 40 3020 • 1.0E+04]

O.OE+00ß • 8.0E+03t Figur•8. •=••øOE'•'031 •. (b) 4.0E+03 are hydrogencomponent of the CME/ENA, G.: "recom,"re- combination;Contributionsof"isg,"different charge processesexchange areoninterstellar shown neu-he • 0.0E+002.0E+031 J . trals; "dust,"charge exchange on dust-producedneutrals. 4.o 5.0 s.o 7.o The startof the geomagneticstorm is markedby the arrow. energy,keV/amu

2.0E+04 t • genCME/ENA flux for 0 = 0 at the initialstage (t < 1.8E+04- 30) of the ENA flux enhancement. The NSW/CME fitix 6 '-- 1.6E+04- dueto chargeexchange on the ISG becomesmore hnportant later (t > 30). For 0 = 180ø the ISG contribution >e 1.4E+04- dominatesfrom the very beginningdue to enhancedmmfl, er •' 1.2E+04- densityof interstellarhelium in the downwindregion. •n 1.0E+04- (b) Chargeexchange on interstellarhelium is especiallyhnpor- ,•- 8.0E+03- tentfor earlyconting CME/ENAs. Obviously,the first to < arriveare ENAs born close to theSun and characterized by E 6.0E+03 highervelocity (energy), and charge exchange cross sections >• 4.0E+03 of protonson heliumatoms rapidly increases with velocity (Figure5). The gravitationalfocusing of heliumexplains 2.0E+03 alsothe large signal of thehelium CME/ENA componentfor 0.0E+00 O = 180ø. A contribution of the recmnbination in the 4.0 5.0 6.0 7.0 8.0 expandingCME plasmaremains relatively small. energy, keV/amu Energydistributions of hydrogenCME/ENAs are shown in Figures 9a and 9b for 0 = 0ø and 180ø for thne Figure9. Energydistributions of hydrogenCME/ENAs at momentst = 20, 30, 40, 50, and60. The energyof the 1 AU at tinges20, 30, 40, 50, and60 for (a) 0 = 0 and particlesarriving at t = 20 and 30 is higherthan the (b) 0 = 180ø;1 tingestep is 300 s. Energyof quiescent energyof ENAs at t = 50 and 60. Curve shapesin NSW atoms is 1.3 keV. GRUNTMAN: NEUTRALSOLAR WIND - GEOMAGNETIc'STORMWARNING 19,225

8.0E+02 Theadvance warning of theapproaching CMEs discussed t • 2O hereis basedOn the ability to measureweak ENA fluxesin ,, 7.0E+02. space.This is a challengingtask and consideratibn of the < 6.0E+02. availabletechniques is out of scopeof thispaper ,as well as of thisjournal. As we mentionedearlier, experhnental A 5.0E+02' techniquesto measureENAs in the energyrange of interest (0.5keV < E < 20 keV)were under gradual development • 4.0E+02 sincelate 1960s[e.g., Waxand Bernstein,1967; Benrvtein et al., 1969; Gruntman amt Morozov, 1982; McEntire atut •" 3.0E+02 E Mitchell; 1989; McComas et al. , 1989; Hsieh et aL , 1992; ? 2.0E+02 Gruntman,1993; Funstenet al. 1993] With r•ently in- creasedinterest and attenti6ndue to prospectsof their r- 1.0E+02 applicationsfor global magnetosphereimaging [e.g.,

O.OE+00 Williams et al. , 1992]. 4.0 5.0 6.0 7.0 8.0 Two majorapproaches for ENA detectionare basedon energy, keV/amu (1) ENA strippingin thinfoils and (2) direcidetection using 6.0E+03 coincidencerexluirements. Thelatter approach hnPleme.0ted t • in the NSW experiment[Gruntman and Morozov, 1982; •' $.0E+03. Grumman and Leonas, 1983; Gruntmane• al., 1990] is optimizedfor conditionsof the quiescentNSW characterized ( by relatively low ENA velocities (--1 keV/anm). Such ..• 4.0E+03' velocitiesresult correspondinglyin a relativelylaygo aberrationangle (for an observermoving with the Earth • 3.0E+03' (b around the Sun) and allows one to aclfieve satisfacto• suppressionOfthe direct solar light by a caretiffdesign of (' 2.0E+O3- the instmment's baffle. 'The CME/ENA flux has much E highervelocity (3-8 keV/amu)and correspondingly muc,b • i.0E+03- smalleraberration angle, which makes it virtuallyimpossibl e tobuild an adequate baffle to protects thesensor fr0h• •ih• solarlight. The t•hnique basedon the ENA stripping• O.OE+00 4.0 5.0 6.0 7.0 8.0 thin foil [e.g., Bernsteinet al., 1969; McContaset al., energy, keV/amu 1991] is preferableunder such conditions. TP e stripping efficiencyof high-pnergyCME/ENAs would be ratherPigh Figure 10. Energydistributions of heliumCME/ENAs at (-- 0.1 for hydrogenatoms and 0.05-0.08 for helitun 1 AU at times 20, 30, 40, 50, and60 for (a) 0 = 0 and atoms).The ions (stripped ENAs) with energies les• than 2- (b) O = 1807; 1 time stepis 300 s. Energyof quiescent 3 keV/amu(corresponding to the quiescentsolar wind NSW atoms is 1.3 keV. precedingthe CME) can be easily rejectedby retarding potentialfilter• andthe remaininghigher-energy ions can be convenientlydeflected (to suppressthe effect of theback- 6. AdvanceStorm Warning groundEuV/UV light),energy analyzed (rexluired energy resolutionE/AE = 10-15),and reliably detected. Develop- The significant "high-energy"(E > 3 keV/amu) mentof efficientdiffraction filters [Gruntman, 1991; Scime CME/ENA flux is expected 2.5 - 3 hours before the et al., 1993]can facilitate the EUV radiationSuppression. arrival of the CME (Figure 8). The advancewarning To conclude,the progressin developmentof ENA instrumentalpackage could be deployedat a spacecrafteither measurementtechnique suggests that both the quiescent insideor outsidethe magnetosphere(say at the Lagrangian NSW and CME/ENAs can be reliablymeasured at 1 AU pointL1) and it shouldinclude ENA, plasma,and magnetic fromthe Sun. Suchmeasurements, performed continuously, field instruments.A detectionof the high-energyENA flux canprovide a basisfor advanceWarning of largegeomagnet- triggersa prelhninaryalert signal and starts the procedure to ic storms,and predictionof their m_agnimdeas well as for assessrecently accumulated and current plasma and nmgnetic remotestudy of thedynamics of CME propagationoutward field measurements. from the Sun. It waspointed out [e.g., Tsurutaniet al., 1992a;Gosling, 1993a]that favorable conditions for developmentof a large Acknowledgement. The Editor thanksJ.D.Mihalov and geomagneticstorm include high differential velocity between anotherreferee for theirassistance in evaluatingthis paper. the CME andsurrounding solar wind and 10ng-duration(as longas daysprior the currentmoment) precursor southward References interplanetarymagnetic field. If the evaluationof the availableplasma and magneticfield •neasure•nentsoutsk!e Ajello, LM., A.I. Stewart,G.E. Thomas,and A. Graps, Solar themagnetosphere shows that these conditions are met, then cycle study of interplanetaryLyman-alpha variations: an alert signal, as well as an estimateof the expected PIONEER orbiter sky backgroundresults, Astro- storm's •nagnitude,are sent to concernedparties (electric phys.J., 317, 964-986, 1987. powergrids, communications, operations, etc.) that Akasofu,S.-I., A sourceof the energyfor geomagneticstorms triggersexecution of protectiveactions by users. and auroras, . SpaceSci., 12, 801-833, 1964a. 19,226 GRUNTMAN: NEUTRAL SOLAR WIND - GEOMAGNETIC STORM WARNING

Akasofu, S.-I., The neutralhydrogen flux in the solar plasma Funsten,H.O., D.J. McComas,and E.E. Scime, Low energy flow,- I, Planet. SpaceSci., 12, 905-913, 1964b. neutral atom imaging techniques, in Instrumentationfor Axford, W.I., The interaction of the solar wind with the MagnetosphericImagery II, Proc. SPIE Int. Soc.Opt. Eng., interstellarmedium, in SolarWind, NASASpec. Publ., SP- 2008, 93-104, 1993. 308, 609-660, 1973. Gloeckler, G., J. Geiss,H. Balsiger, L.A. Fisk, A.B. Galvin, Banks,P.M., Interplanetaryhydrogen and heliumfrom cosmic F.M. Ipavich, K.W. Ogilvie, R. yon Steiger,and B. Wilken, dustand the solarwind, J. Geophys.Res., 76, 4341-4348, Detection of interstellar pick-up hydrogen in the solar 1971. system, Science,261, 70-73, 1993. Baranov,V.B., Gasdynamicsof the solarwind interactionwith Gonzales,W.D., and B.T. Tsurutani, Terrestrialresponse to the interstellarmedium, SpaceSci. Rev., 52, 89-120, 1990. eruptivesolar flares' Geomagneticstorms, in EruptiveSolar' Baranov, V.B., and Yu.G. Malama, Model of the solar wind Flares, editedby Z. Svestka,B.V. Jackson,and M.E. Ma- interaction with the local interstellar medium: Numerical chado, pp.277-286, Springer-Verlag,New York, 1992. solutionof self-consistentproblem, J. Geophys.Res., 98, Gosling,J.T., In situobservations of coronalmass ejections in 15,157-15,163, 1993. interplanetaryspace, in EruptiveSolar' Flares, editedby Z. Barnett, C.F., H.T. Hunter, M.I. Kirkpatrick, I. Alvarez, C. Svestka, B.V. Jackson,and M.E. Machado, pp.258-267, Cisneros,and R.A. Phaneuf, Atomic data for fusion. Colli- Springer-Verlag,New York, 1992. sionsof H, H2, He, and Li atomsand ions with atomsand Gosling,J.T., Coronalmass ejections: The link betweensolar molecules, Oak Ridge Natl. Lab., ORNL-6086/V1, Oak and geomagneticactivity, Phys. Fluids B, 5, 2638-2645, Ridge, Tenn., 1990. 1993a. Bernstein,W., R.L. Wax, N.L. Sanders,and G.T. Inouye, An Gosling, J.T., The solar flare myth, J. Geophys.Res., 98, energy spectrometer for energetic (1-25 keV) neutral 18,937-18,949, 1993b. hydrogen atoms, in Small Rocket InstrumentationTech- Gosling, J.T., D.J. McComas, J.L. Phillips, and S.J. Bame, niques, pp.224-231, North-Holland, New York, 1969. Geomagneticactivity associatedwith Earth passageof Bertaux, J.-L, Helium and hydrogenof the local interstellar interplanetaryshock disturbances and coronal mass ejections, mediumobserved in the vicinity of the Sun, in IAU Collo- J. Geophys.Res., 96, 7831-7839, 1991. quium No.81, NASA Conf. Publ., CP 2345, 3-23, 1984. Gruntman,M.A., Neutral componentof the solarwind at the Bertaux J.-L., R. Lallement, V.G. Kurt, and E.N. Mironova, Earth's orbit (in Russian), Cosmic Research, 18, 649-651, Characteristicsof the local interstellarhydrogen determined 1980. from PROGNOZ 5 and6 interplanetaryLyman c• line profile Gruntman,M.A., The effectof neutralcomponent of the solar measurementswith a hydrogen absorptioncell, Astron. wind on the interactionof the solarsystem with the interstel- Astrophys., 150, 1-20, 1985. lar gas flow, Sov. Astron. Lett., 8, 24-26, 1982. Bertin, P., R. Lallement,R. Ferlet, and A. Vidal-Madjar, Gruntman,M.A., Two-stepphotoionization of hydrogenatoms Detection of the local interstellar cloud high-resolution in interplanetaryspace, Planet. SpaceSci., 38, 1225-1230, spectroscopyof nearbystars: Inferenceon the heliospheric 1990. interface, J. Geophys.Res., 98, 15,193-15,197, 1993. Bleszynski,S., S. Grzedzielski,D. Rucinski,and J. Jakimiec, Gruntman,M.A., Submicronstructures: promising filters in Expectedfluxes of --• 1 keV neutralH atomsin interplane- EUV A review, in EUV, X-Ray, and Gamma-Ray tary space; comparisonwith the u.v. backgroundand Instrumentationfor Astronomy,Proc. SPIE htt. Soc. Opt. Eng., 1549, 385-394, 1991. possibilityof detection, Planet. SpaceSci., 40, 1525-1533, 1992. Gruntman,M.A., Anisotropyof the energeticneutral atom flux Blum, P., and H.J. Fahr, Interaction between interstellar in the heliosphere,Planet. SpaceSci., 40, 439-445, 1992. Gruntman,M.A., A new techniquefor in situ measurementof hydrogenand the solar wind, Astron. Astrophys.,4, 280- 290, 1970. the compositionof neutral gas in interplanetaryspace, Dalgarno, A., and A.C. Allison, Photodissociationof molecu- Planet. SpaceSci., 41,307-319, 1993. lar hydrogenon Venus, J. Geophys.Res., 74, 4178-4180, Gruntman, M.A., and V.A. Morozov, H atom detection and 1969. energyanalysis by useof thin foils and TOF technique,J. Dessler, A.J., Solar wind and interplanetarymagnetic field, Phys. E, 15, 1356-1359, 1982. Rev. Geophys.,5, 1-41, 1967. Gruntman, M.A., and V.B.Leonas, Neutral Solar Wind. Fahr, H.J., On the influence of neutral interstellar matter on Possibilityof ExperimentalStudy, Report (PreptinO the upper , Astrophys.Space Sci., 2, 474-495, No.825, SpaceResearch Institute (IKI) of the USSRAcade- 1968a. my of Sciences,Moscow, 1983. Fahr, H.J., Neutralcorpuscular energy flux by charge-transfer Gruntman,M.A., S. Grzedzielski,and V.B. Leonas, Neutral collisionsin the vicinityof the sun, Astrophys.Space Sci., solarwind experiment, in Physicsof theOuter Heliosphere, 2, 496-503, 1968b. edited by S. Grzedzielskiand D.E. Page, pp.355-358, Fahr, H.J., The extraterrestrialUV-background and the nearby Pergamon, New York, 1990. interstellarmedium, SpaceSci. Rev., 15, 483-540, 1974. Hall, D.T., D.E. Shemansky,D.L. Judge,P. Gangopadhyay, Fahr, H.J., H.W. Ripken,and G. Lay, Inter-Zodiak: A high- andM.A. Gruntman, Heliospherichydrogen beyond 15 AU: apogee rocket experiment for the observationof dust- evidencefor a terminationshock, J. Geophys.Res., 98, generatedneutrals in the solar vicinity, Proceedingsof the 15,185-15,192, 1993. Vth ESA-PAC Symposiumon EuropeanRocket and Balloon Han, S.M., S.T. Wu, M. Dryer, A three-dimensional,time- Programmesand Related Research, European Space Agency dependentnumerical modelling of super-sonic,super-alfvenic Spec. Publ., SP-152, 449-551, 1980. MHD flow, Cornput.Fluids, 16, 81-103, 1988. Fahr, H.J., H.W. Ripken, and G. Lay, Plasma- dustinterac Hildner E., Characteristicsof coronal mass ejections, in tion in the solarvicinity and their observationalconsequenc- Eruptive Solar Flares, edited by Z. Svestka,B.V. Jackson, es, Astron.Astrophys., 102, 359-370, 1981. and M.E. Machado, pp.227-233, Springer-Verlag,New Fahr, H.J., Filtration of the interstellar neutrals at the helio York, 1992. sphericinterface and their couplingto the solar wind, in Holzer, T.E., Neutral hydrogenin interplanetaryspace, Rev. Physicsof the OuterHeliosphere, edited by S. Grzedzielski Geophys.,15, 467-490, 1977. and D.E. Page, pp.327-343, Pergamon,New York, 1990. Hsieh, K.C., C.C. Curtis, C.Y. Fan, and M.A. Gruntman, GRUNTMAN: NEUTRAL SOLAR WIND - GEOMAGNETIC STORM WARNING 19,227

Techniquesfor the remotesensing of spaceplasma in the ionization on the distribution of interstellar helium in the heliospherevia energeticneutral atoms: A review, in Solar inner heliosphere;possible consequences for determination WindSeven, editedby E. Marschand R. Schwenn,pp.357- of interstellarhelium parameters, Astron. Astrophys.,224, 364, Pergamon,New York, 1991. 290-298, 1989. Hsieh, K.C., K.L. Shih, D.J. McComas, S.T. Wu, and Scime, E.E., H.O. Funsten, D.J. McComas, K.R. Moore, and C.C. Wu, Forecastingthe arrival of fast coronalmass ejecta M.A. Gruntman, Advances in low energy neutral atom at Earth by the detectionof 2-20 keV neutral atoms, in imaging techniques, Instrumentationfor Magnetospheric Instrumentationfor MagnetosphericImagery, Proc. SPIE !nt. Imagery lI, Proc. SPIE lnt. Soc. Opt. Eng., 2008, 74-82, Soc. Opt. Eng., 1744, 72-78, 1992. 1993. Hundhausen, A.J., Sizes and locations of coronal mass SchwennR., H. Rosenbauer,and K.-H. Muhlhauser, Singly- ejections:SMM observationsfrom 1980 and 1984-1989, J. ionized helium in the driver gas of interplanetaryshock Geophys.Res., 98, 13,177-13,200, 1993. wave, Geophys.Res. Lett., 7, 201-204, 1980. Illing, R.M.E., and E. Hildner, Neutralhydrogen in the solar Shemansky,D.E. andD.T. Hall, The distributionof atomic wind at 1 AU, Eos Trans. AGU, 75(16), Spring Meeting hydrogenin the magnetosphereof , J. Geophys.Res., suppl., 261, 1994. 97, 4143-4161, 1992. Jackson,B., R. Gold, andR. Altrock, The solarmass ejection Sistoe, G.L., and N.R. Mukherjee, Upper limits on the Lunar imager, Adv. SpaceRes., 11,377-381, 1991. atmospheredetermined from solar-windmeasurements, J. Lallement, R., J.-L. Bertaux, E. Chassefiere,and B.R. Sandel, Geophys.Res., 77, 6042-6051, 1972. Lyman-alphaobservations from Voyager (1-18 AU), in Thomas, G.E., The interstellar wind and its influence on the Physicsof the OuterHeliosphere, edited by S. Grzedzielski interplanetaryenvironment, Ann. Rev. Earth Planet. Sci., and D.E. Page, pp.73-82, Pergamon,New York, 1990. 6, 173-204, 1978. Lallement, R., J.-L. Bertaux, and J.T. Clarke, Decelerationof Tsurutani, B.T., and W.D. Gonzalez, On the solar and inter interstellarhydrogen at the heliosphericinterf•tce, Science, planetarycauses of geomagneticstorms, Phys. FluidsB, 5, 260, 1095-1098, 1993. 2623-2630, 1993. Lanzerotti, L.J., Comment on "Great magneticstorms" by Tsurutani, B.T., W.D. Gonzalez, F. Tang, and Y.T. Lee, Tsurutaniet al., Geophys.Res. Lett., 19, 1991-1992, 1992. Great magneticstorms, Geophys.Res. Lett., 19, 73-76, Lay, G., H.J. Fahr, and H.U. Nass, INTERZODIAK II: 1992a. Observationof EUV-resonanee radiation, Proceedingsof Tsurutani, B.T., W.D. Gonzales, F. Tang, Y. T.Lee, M. Ninth ESA/PAC Symp. European Rocket and Balloon Okada, and D. Park, Reply to L.J.Lanzerotti: Solar wind Programmesand Related Research, Europeart Space Agency correctionsand an estimationof the efficiencyof viscous Spec. Publ., SP-291, 173-176, 1989. interaction, Geophys.Res. Lett., 19, 1993-1994, 1992b. Lay, G., H.U. Nass,and H.J. Fahr, Contaminationof terrestri- Wallis, M., Local interstellarmedium, Nature, 254, 202-203, al EUV observationsby energeticparticles, Proceedingsof 1975. !0th ESA Symp.European Rocket and BalloonProgrammes Wax, R.L., and W. Bernstein, Energyindependent detector for andRelated Research, European Space Agency Spec. Publ., total hydrogenfluxes in the range 1-10 keV for spaceand SP-317, 339-342, 1991. laboratoryapplications, Rev. Sci. Inst•,t., 38, 1612-1615, McComas,D.L., B.L. Barraclough,R.C. Elphic, H.O. Funsten 1967. • III, and M.F. Tomsen, Magnetosphericimaging with low- Williams, D.J., E.C. Roelof, and D.G. Mitchell, Global energyneutral atoms, Proc. Natl. ,4cad. Sci. U.S.A., 88, magnetosphereimaging, Rev. Geophys.,30, 183-208, 1992. 9598-9602, 1991. witte, M., H. Rosenbauer,M. Banaszkiewicz,and H. Fahr, McEntire, R.W., and D.G. Mitchell, Instrumentationfor The Ulyssesneutral gas experiment: Determinationof the global magnetosphereimaging via energeticneutral atoms, velocity and temperatureof the interstellarneutral helium, in Solar SystemPlasma , Geophys. Monogr. Ser., Adv. SpaceRes., i3(6), 121-130, 1993. edited by J.H. Waite Jr., J.L. Burch, and R.L. Moore, Wu, F.M., and D.L. Judge, Temperatureand flow velocityof pp.69-80,AGU, Washington,D.C., 1989. the interplanetarygases along solar radii, Astrophys.J., Meier, R.R., Someoptical and kinetic properties of the nearby 231,594-605, 1979. interstellargas, Astron.Astrophys., 55, 211-219, 1977. Patterson, T.N.L., F.S. Johnson, and W.B. Hanson, The M.A.Gruntman, Departmentof AerospaceEngineering, distributionof interplanetaryhydrogen, Planet. SpaceSci., MC-1191, Universityof SouthernCalifornia, Los Angeles, CA ./1,767-778, 1963. 90089-1191. (e-mail: mikeg•spock.use.edu) Ripken, H.W., and H.J. Fahr, Modification of the local interstellar gas properties in the hellosphericinterface, Astron.Astrophys. , 122, 181-192, 1983. (ReceivedMarch 25, 1994; revisedJune 7, 1994; Rucinski,D., andH.J. Fahr, The influenceof electronimpact acceptedJune 15, 1994.)