Radio Emission from the Heliopause Triggered by an Interplanetary Shock Author(S): D

Home , Bow shock

Radio Emission from the Heliopause Triggered by an Interplanetary Shock Author(s): D. A. Gurnett, W. S. Kurth, S. C. Allendorf, R. L. Poynter Source: Science, New Series, Vol. 262, No. 5131 (Oct. 8, 1993), pp. 199-203 Published by: American Association for the Advancement of Science Stable URL: http://www.jstor.org/stable/2882326 Accessed: 16/06/2010 15:06

Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use.

Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at http://www.jstor.org/action/showPublisher?publisherCode=aaas.

Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission.

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

American Association for the Advancement of Science is collaborating with JSTOR to digitize, preserve and extend access to Science.

http://www.jstor.org ARTICLE....- Radio Emission from the Heliopause Triggered by an InterplanetaryShock

D. A. Gurnett,W. S. Kurth,S. C. Allendorf,R. L. Poynter

A strong heliospheric radio emission event has been detected by Voyagers 1 and 2 in the locatedat R = 39.0 AU, e = -11.7?, and frequency range of 2 to 3 kilohertz. This event started in July 1992 and is believed to have X = 2830. The spectrogramscover a fre- been generated at or near the heliopause by an interplanetary shock that originated during quency range from 1 to 4 kHz and a time a period of intense solar activity in late May and early June 1991. This shock produced large spanof 1 year, fromday 120 of 1992 to day plasma disturbances and decreases in cosmic ray intensity at Earth, Pioneers 10 and 1 1, 120 of 1993. and Voyagers 1 and 2. The average propagation speed estimated from these effects is 600 Two primaryspectral components can to 800 kilometers per second. After correction for the expected decrease in the shock speed be seen in Fig. 1: a main emissionband at in the outer heliosphere, the distance to the heliopause is estimated to be between 1 16 about 2.0 kHz, and a seriesof narrowband and 177 astronomical units. emissionsdrifting upward in frequencyat a rate of about 3.0 kHz year-1, reaching a peakfrequency of about3.6 kHz. The main band has a sharplydefined low-frequency The heliopauseis the boundarybetween a was locatedat a heliocentricradial distance cutoff at 1.8 kHz. Both componentsmust hot (- 105 K) ionized gas flowingoutward of R = 50.8 AU and a solarecliptic latitude consist of electromagneticwaves because fromthe sun, called the solarwind (1), and and longitudeof 1B= 33.60 and A = 2450 they occur at frequencieswell above the the interstellarmedium, which is a relative- (as of 1 January1993), and Voyager2 was local plasma frequency. The plasma fre- ly cool (- 104 K), partially ionized gas between the stars.Because the sun is mov- ing with respect to the nearbyinterstellar Intensity(dB) medium,the heliopauseis expectedto form 0 10 a teardrop-shapedsurface, the nose of l I which is towardthe directionof arrivalof Voyager1 the interstellargas (2). The region inside the heliopause is called the heliosphere. 4.0 Estimatesof the distanceto the heliopause based on our limited knowledge of the interstellarmedium have variedfrom a few 3.0 tens to severalhundred astronomical units (1 AU = 1.49 x 108 km). For a recent review of the heliosphereand its interaction with the interstellarmedium, see Suess 2.0 (3). N Fourspacecraft, Pioneers 10 and 11 and I Voyagers1 and 2, areon escapetrajectories 1.0 fromthe sun to studythe outer heliosphere and to penetrateinto the interstellarmedium. None has yet reachedthe heliopauseor ) Voyager2 the solar wind terminalshock, which is a C-a) 4.0 standingshock that is expected to form in IL. the supersonicsolar wind flowwell insideof the heliopause. The radio emission de- 3.0 scribedin this reportand its interpretation provide a direct measurementof the distance to the heliopause. 2.0 Description of the Event Frequency-timespectrograms of the radio 1.0 emissionevent (Fig. 1) were obtainedfrom the widebandplasma wave system (PWS) 120 180 240 300 360 0 60 120 on Voyager (4), which provides periodic 1992 1993 samplesof the electricfield waveform over a bandwidthof 50 Hz to 10 kHz. Voyager 1 Time (day and year) Fig. 1. Frequency-time spectrograms of the 1992-93 heliospheric radio emission event from (top) D. A. W. and in Gurnett, S. Kurth, S. C. Allendorf are the the Department of Physics and Astronomy, University Voyager 1 and (bottom) Voyager 2. The color indicates the electric field intensity, with red being of Iowa, Iowa City, IA 52242. R. L. Poynter is at the Jet most intense and blue being the least intense. Sampling times are indicated by marks at the top of Propulsion Laboratory, Pasadena, CA 91109. each panel.

SCIENCE * VOL. 262 * 8 OCTOBER 1993 199 quency (f = 9000 VN Hz, whereN is the Day electroncfensity in electronsper cubic cen- = 8.0 _Complete rolls 3.11 kHz Roll-axis 311 (lookingtoward timeter)is the low-frequencycutoff of free- E Earth) , - _ 036 Source space electromagnetic waves and corre- __ 1 / _- directions 6.0- spondsapproximately to the high-frequency - - -, (3.1 kHz) limit of locallygenerated plasma waves. For _ _, f ~~~220 the time period of interest, the plasma p4.0- ' Sun's velocity relativeto frequency,as determinedby the Voyager2 - Voyager I nearby Fig. - interstellarmedium plasmainstrument, ranged from 0.4 to 1.3 Day220 kHz, with a mean of around0.7 kHz (5). 7 Aug 1992 Fig. 3. Source directions at 3.11 kHz deter- mined from roll maneuvers on days 220 and The exact startingtime of the radio emis- 'L 18 20 22 24 0 Time (hours) 311, 1992, and on day 36, 1993. The view is sion is difficultto determinefrom the wide- from the spacecraft looking along the roll axis A banddata because the time resolutionearly Fig. 2. plot of the electric field intensity in the (, = -33.7?, X = 63.70) toward the Earth. The in the event is only one spectrumper week 3.11 -kHz channel during a 10-turn roll maneu- horizontal (reference) direction has been taken (Fig. 1). A more accuratestarting time can ver on day 220, 1992 (11-point sliding average, to be the direction of the sun's motion with be obtained from the PWS 16-channel best fit marked by smooth curve). A clear respect to the nearby interstellar medium (p = roll-modulation signal can be seen. The phase 5.00, X = 2540), as given by Ajello et al. (24). spectrum-analyzerdata, which providesone of the roll modulation gives the direction to the measurementevery 16 s. These data show source. that the radioemission first appeared in the 1.78-kHz channel on day 188 (6 July) of consistent with a source location near ei- 1992. After the start of the event, the frequencyrange to detect the radio emis- ther the nose or the tail of the heliosphere. intensity graduallyincreased over a period sion. No roll-modulationeffects were ob- For the specific geometry involved, the of months, reachinga peakin earlyDecem- servedin the 1.78-kHzchannel. However, locationnear the nose or tail turnsout to be ber 1992. Thereafter,the intensity gradu- a clear roll-modulationsignal was observed nearly independentof the distance to the ally decreasedand was almostdown to the in the 3.11-kHz channel during all three source (<10 variationfor distancesgreater receiver-noiselevel as of July 1993. maneuvers(Fig. 2). The sinusoidalmodu- than 100 AU). The variationsin the ob- A strikingcharacteristic of the overall lation in the electricfield intensityat twice served source directionswere probablyre- event is the similarityof the spectrumsat the rotation period of the spacecraftis lated to the fact that the 3.1 1-kHz channel the two spacecraft,even though they are clearly evident. To determine the ampli- was respondingto the upward-driftingnar- separatedby 44.6 AU. This similaritysug- tude and phase of the modulation,we fit a rowbandfeatures (see Fig. 1), which were gests that the source is at a considerable sine function to the data using a least- clearly evolving during the course of the distance, greater than 44 AU. From the squares fitting procedure. All three roll event. maximum radiation intensity (1.8 x 10-17 maneuvershad similar modulation signa- W m-2 Hz-'), the radiated power is esti- tures.The peak-to-peakmodulation ampli- Relation to the Great Forbush mated to be at least 1013 W. This radio tudes range from about 10 to 20%. These Decrease of 1991 source is much strongerthan any known relativelylow modulationamplitudes indi- planetaryradio source and is probablythe cate that the sourceis eitherrelatively large Only one previousheliospheric radio emis- most powerfulradio emission in the solar (>600) or is locatedwell awayfrom the roll sion event has been observedwith intensi- system. axis (.40?), or a combination of these ties comparablewith those of the 1992-93 effects. The existence of a significantan- event. This event occurredin 1983-84 and Radio Direction Finding isotropyalso indicatesthat the radiation(at was firstdescribed by Kurthet al. (7). Five Measurements 3.11 kHz) is not trappedin a high-Q cavity other extremely weak events have been as suggestedby Czechowskiand Grzedziel- observedbetween these two major events Voyager is normally stabilizedin a fixed ski (6): Multiple reflections in a cavity (8), one in late 1985, one in 1989, and inertial orientationwith the high-gainan- would be expected to quickly lead to an three in 1990-91. Several investigators tenna pointed toward Earth. However, isotropicelectric field distribution. have searched for unusual effects in the once every 3 months, the spacecraftper- For a rotatingelectric dipole, it can be solar wind that might triggerthese radio formsa seriesof rolls aroundthe axis of the shown that the sourcemust lie in a plane emissions. For example, McNutt (9) first high-gainantenna to calibratethe magne- perpendicularto the antenna axis at the suggested that a high-speed solar wind tometer. Three such roll maneuverswere time of maximumsignal strength, which stream-could trigger the radio emission performedduring the emission event, the can be determinedfrom the phase of the when it reachedthe terminalshock. This firsttwo on days 220 and 311 of 1992 and roll modulation. If the antenna axis is idea was explored furtherby Grzedzielski the third on day 36 of 1993. Because the perpendicularto the roll axis, as it is during and Lazarus(10), who identifieda seriesof PWS dipole antenna axis is perpendicular these maneuvers,then the plane through dynamicpressure increases in the solarwind to the roll axis, these maneuverscan be the source also contains the roll axis. To that they believed were responsiblefor the usedto performradio direction finding mea- visualizepossible source locations, it is con- 1983-84, 1985, and 1989 events. Despite surements.The 16-channel spectrum-ana- venient to construct a diagram looking the possiblemerit of these suggestions,the lyzerdata must be used for these measure- along the roll axis with all vectorsprojected cause-effectrelationship was not convinc- ments becausethe samplerate of the wide- onto a plane perpendicularto the roll axis ingly demonstrated,and other sourcescon- band spectrumis too low to resolvethe roll (Fig. 3). In such a diagram, the plane tinued to be considered.For an overviewof modulation.Also, becauseof a data system through the source reducesto a line with the situationbefore the 1992-93 event, see failureon Voyager 2 shortly after launch, two possible directions, as shown by the Kurth(8). directionfinding measurements can only be dashedlines with arrowsat either end. Because heliospheric radio emission pefformed on Voyager 1. The observedsource directionstend to events comparablewith the 1992-93 event Only two spectrum analyzer channels, cluster around the projected direction of are extremelyrare, one would expect that 1.78 and 3.11 kHz, are in the proper the sun'smotion (Fig. 3). The clusteringis the solar wind trigger would also be an

200 SCIENCE * VOL. 262 * 8 OCTOBER 1993 ARTICLE

Solarflares datafrom the Deep Riverneutron monitor, radioemission is producedby an interplan- IA A600 la I X I I I{ which recordsthe cosmic ray intensity at etaryshock at the plasmafrequency (fp), its Frorbush the Earth(Fig. 4A). This decreasewas the harmonic(2fp), or both. The generationof 500 _ idecreasf deepestdepression in the cosmic rayinten- radioemission at f and 2fpby an interplan- Deep River sity ever observedby a neutronmonitor in 400_ neutronmonitor- etary shock is weli known and is the basic over 30 years of observations.Cosmic ray mechanisminvolved in the generationof B 0.7 Pioneer1 0 Forbush intensities from Pioneer 10 and 11, at 53 type II solarradio bursts (19). It is also the 8sE 03 8 .o,decrease and 34 AU (Fig. 4B), and fromVoyagers 1 only mechanismthat has been previously - 53 AU and 2, at 46 and 35 AU (Fig. 4C), were consideredas an explanationfor the helio- Pioneer 1 Shock Forbush also recorded.As the shocks generatedby sphericradio emissions (20). _0.5 the solar activity propagatedoutward from Three boundariesmust be consideredas _ :Lir~~''/'"A4 7decrease the sun, they arebelieved to have coalesced the interplanetaryshock propagatesout- 260 7f / v 34 AU c into a single shock followedby a region of wardfrom the sun: the terminalshock, the ~~0.3 turbulent high-speed plasma called a heliopause, and the bow shock (Fig. 5). ShockN merged interaction region (16). As the Most previousheliospheric radio emission C 340- Forbush disturbedplasma swept over the Pioneer modelshave focusedon the terminalshock and Voyager spacecraft, roughly 3 to 4 as the sourceregion. However,in this case, months after the flare activity, strong de- the radio emissioncannot be generatedat 20o0 20300 00 l 0 creasesoccurred in the cosmic ray intensi- the terminalshock or in the regionbetween ties, firstat Pioneer 11, then at Voyagers2 the terminalshock and the heliopause.The and 1, and finallyat Pioneer10. The shock reasonis that the electronplasma frequency itself was also detected, firstby the magne- is too low. Given the observedpropagation tometer on Pioneer 11 (17), then by the speedof 600 to 800 km s1- and traveltime plasmainstrument on Voyager2 (5), and of -1.1 years, the terminal shock would 1991 1992 finally by the plasmawave instrumenton have to be located at a distance of 140 to Time(day and year) Voyager1 (14) (indicatedby arrowsin Fig. 186 AU. At this great distance, the elec- Fig. 4. Cosmic ray counting rates at (A) Earth, 4). Fromthe timing of the variousevents, tron plasmafrequency, which variesas 1/R, (B) Pioneers 10 and 11 (53 and 34 AU),and (C) Van Allen and Fillius estimate that the would be only about 100 to 150 Hz. Be- Voyagers 1 and 2 (46 and 35 AU). The sharp averagepropagation speed is 820 km s-5 cause the electron plasma frequencycan decreases in the cosmic raycounting rates are Webber and Lockwood give propagation only increaseby a factorof 2 at a shock and produced by an outward-propagatinginter- 1. planetaryshock generated by a series of large speedsranging from 600 to 800 km s- The because the highest frequencythat can be solar flares in late May and early June 1991. overallpropagation time, fromthe onset of generatedis twice the electron plasmafre- [Adaptedfrom (14) and (15)]. the solaractivity on day 145, 1991, to the quency, it would be impossibleto generate onset of the radio emission event on day the observedfrequencies. 188, 1992, is 408 days or 1.12 years. At the heliopause,the situationis much unusual event. During the period preceding Although the 1992-93 radio emission better. The plasma density at the helio- the 1992-93 radio emission event, there is event and the interplanetaryshock associ- pause is controlled by pressurebalance. one event, the great Forbush decrease of ated with the 1991 Forbushdecrease are Becausethe temperatureon the interstellar 1991, that fits this requirement perfectly. both extraordinaryevents, one case does side of the heliopause is expected to be For many years it has been known that the not prove a cause-effectrelationship. How- much colder than on the solar wind side, sun occasionally ejects an energetic burst of ever, upon reexaminationof the 1983-84 the electronplasma frequency can increase plasma called a coronal mass ejection (11). heliosphericradio emission event, which by a large factor at the heliopause, to a The ejection is often associated with a started on day 242, 1983, a very large value that is comparablewith the plasma series of solar flares and produces an inter- (21%) Forbushdecrease was found a little frequencyin the interstellarmediumn; From planetary shock wave that moves outward over 1 year earlier,on day 195, 1982 (18). numerical simulations (21) and various from the sun at high speed, up to 1000 km The time interval between this Forbush physical arguments,the plasma-frequency s-5. The shock wave is typically followed by event and the onset of the 1983-84 radio profilein the vicinity of the heliopauseis a turbulent high-speed stream of plasma. As emissionis 412 days, or 1.13 years, almost believed to be as shown in Fig. 6. The the shock propagates outward through the exactly the same as the 1992-93 radio plasmafrequency immediately inside of the heliosphere, the turbulent magnetic fields emission event. Also, the propagation bow shock dependson the strengthof the embedded in the high-speed stream inhibit speeds,810 and 850 km s-1, given by Van bow shock and cannot be more than twice the entry of cosmic rays, causing a tempo- Allen and Randall and by Webber et al. the plasma frequency in the interstellar- rary decrease in the cosmic ray intensity. (18), are almost exactly the same as the medium. Remote sensing measurements This effect is called a Forbush decrease speed given by Van Allen and Fillius (14) (22) suggestthat the electrondensity in the (12). for the 1991 Forbushevent. Thus, the two local interstellarmedium is about 0.03 to During the period from 25 May to 15 strong heliosphericradio emission events 0.1 cm-3, which corresponds to a plasma June 1991, six major solar flares were ob- observedby Voyagerboth appearto have frequencyof about 1.6 to 2.8 kHz. These served (13). This sequence of intense solar been triggered by interplanetaryshocks plasmafrequencies are in the same general activity produced a series of strong inter- with largeForbush decreases. range as the observedradio emission fre- planetary shocks and one of the largest quencies and are thereforeconsistent with Forbush decreases ever observed. The main Interpretationof the Radio generationin the vicinity of the heliopause. features of this event were first discussed by Emission Spectrum Justwhy the radioemission should turn on Van Allen and Fillius (14) and Webber and as the interplanetaryshock encountersthe Lockwood(15). Shortly after the onset of The frequency-time spectrum of the 1992- heliopause is unknown. Most likely, the the period of intense solar activity, a large 93 radio emission event is very complex. To onset is related to the much lower temper- (-30%) Forbush decrease developed in the interpret the spectrum, we assume that the atures of the interstellar medium, which

SCIENCE * VOL. 262 * 8 OCTOBER 1993 201 wouldreduce the Landaudamping, thereby Fig. 5. A sketch of the possiblycausing higher radio emission in- heliosphere and its an- . Bowshock tensities. ticipated boundaries. Once the interplanetary shock has The heliopause is the Heliopause C crossed the heliopause, the emission fre- boundary between the 'B \ solar wind and the in- quency depends on the plasma frequency Terminalsho terstellar plasma. Be- Temia Axh\ encounteredby the shock as it propagates cause the solar wind is - Inter- throughthe post-heliopauseplasma. Numer- supersonic, a standing planetary Interste\la ical simulationsshow that there is a pile up shock, called the termi- schock V2 A Interstelar of plasmaaround the nose of the heliosphere nal shock, is expected Pio0 A A Pi (21). A cut along the line A-B-C in Fig. 5 to formin the solarwind \ VI / would thereforebe expectedto give a peak flow well inside of the / in the electronplasma frequency profile, as heliopause. A second - Pile-up shown by the curve A-B-C in Fig. 6. A standing shock, called / region the bow shock, is also r_o shockpropagating through this regionwould expected to formin the then give an emission frequencythat in- interstellarplasma flow / creaseswith increasingtime, as indicatedby upstream of the helio- the curvelabeled A-B in the inset to Fig. 6. sphere. The outward The risingemission frequency is believedto propagatinginterplane- accountfor the upward-driftingnarrowband taryshock produced by the solar activityin late Mayand earlyJune 1991 is indicatedby the circle emissionsevident in Fig. 1. This interpreta- withoutward directed arrows.The approximatepositions of Pioneers 10 and 11 (Pl0 and Pl 1) and tion also predictsa sourcenear the nose of Voyagers 1 and 2 (Vl and V2) are indicated by the dots. PointsA, B, and C correspondwith those the heliopause,which is consistentwith the in Fig. 6. direction-findingmeasurements described earlier.It is not clearwhether the radiation is producedat fp, 2fp, or both. Finally, we consider the low-frequency Forthe segmentof the shockfront prop- cutoffof the spectrumat 1.8 kHz. There are /Dr B agating through the flanks of the helio- two possibleinterpretations: first, that it is a Termia BoC ~~~U- ~ B'pu ~ ~ ~ B pause, as along line A-B'-C in Fig. 5, the propagationcutoff at the plasmafrequency, rime emission frequencyshould be nearly con- and second, that it is a characteristicemis- stant. Radiationfrom this regionis believed sion frequency (either or in the ! Vshh - Interstellar fp 2fp) UL. d Interplanetary plasma to account for the main emissionband in source. If it is a propagationcutoff, then it Solarwn shock Fig. 1, which is nearly constant at a fre- most likely correspondsto the plasmafrequency of about 2.0 kHz. In the simplest quency in the post-heliopauseregion well Terminal Heliopause Bow interpretation,the radiationwould not be away from the nose of the heliosphere, shock shock trappedin the heliosphericcavity. Howev- where the plasma density is comparable Fig. 6. A representativeelectron plasma fre- er, the issue of trappingdepends sensitively with the interstellarvalue. The electron quency (fp) profilethrough the terminalshock, the heliopause, and the bow shock. The peak in on the details of the plasma frequency density in this region would then be about the profile along A-B-C (see also Fig. 5) is profile (which is poorly known) and the 0.04 cm-3. If the cutoff is a sourceeffect, caused by the "pile up" of plasma near the exact point where the emissionturns on. It then the plasmadensity in the sourcewould nose of the heliosphere and is believed to is possiblethat some or all of the radiation be 0.04 cm-3 if the emission is at the account for the upward-driftingnarrowband generatedin the immediatevicinity of the fundamental,or 0.01 cm-' if it is at the emissions (Fig. 1). The nearlyconstant frequen- heliopausecould be trappedin the helio- harmonic. cy emission at -2.0 kHz is believed to be spheric cavity. The absence of detectable produced from the flanks of the heliosphere, roll modulation in the 1.78-kHz channel Distance to the Heliopause along A-B'-C, where the plasma frequency is does not rule out trappingin this frequency nearlyconstant. The 1.8-kHzcutoff would then range. We are also not completelycertain The distance to the can be be indicative of the plasma frequency in this heliopause region. (Inset) Frequency-timeprofile. how to interpretthe enhancementat about computed from the propagationspeed of 2.7 kHz. One possibilityis that the main the interplanetaryshock and the travel bandat 2.0 kHzis causedby emissionat the time from the sun to the heliopause.The fundamental, fp, and the enhancement difficultywith this calculation is that the the terminal shock to the distance to the around2.7 kHz is causedby emissionat the shock almost certainly slows down after heliopause. With these two parameters, it harmonic, 2fp. An obvious problemwith passing through the terminal shock. Be- can be shown that the distance -to the this interpretationis that the ratio of the cause the propagationspeed in the region heliopause is given by frequencyof the enhancement to the fre- beyond the terminal shock is unknown, quency of the main band is only 1.35, certain simplifying assumptionsmust be which does not indicate a harmonicrela- made. As a simple model, we assumethat RH =V1T 1 ( ) (1) tion. Similar,though smaller, discrepancies the shock propagateswith a constantspeed also occurfor type II solarradio bursts (19). V1inside of the terminalshock and with a where T is the total travel time from the Another possibility is that the 2.0- and slower constant speed V2 outside of the sun to the heliopause. For an initial esti- 2.7-kHz componentsare emitted from two terminalshock. To take into account the mate, we use T = 1.1 years and V1 = 600 regionsthat have differentplasma densities, unknown speed V2, we introducea speed to 800 km s- 1 These values are consistent hence differentplasma frequencies. With- parameterot = V2/V1.Because the distance with both the 1983-84 and 1992-93 out more detailed information on the plas- to the terminalshock is also unknown,we events. For the shock speed parameter ax, ma density distribution, it is difficult to also introduce a distance parameter8 = numerical simulations (23) suggest a nom- distinguish between such interpretations. RT/RH,which is the ratioof the distanceto inal value of about 0.7, with a range from

202 SCIENCE * VOL. 262 * 8 OCTOBER 1993 ARTICLE

Table 1. The distance to the heliopause (in improve the accuracyof these estimates. (1984); L. F. Burlaga, F. B. McDonald, N. F. Ness, astronomical units) as a function of the param- A. J. Lazarus, ibid. 96, 3780 (1991). 17. D. Winterhalter, E. J. Smith, L. W. Klein, Eos 73, eter at (the ratio of the interplanetary shock REFERENCESAND NOTES speed beyond the terminal shock, V2, to the 237 (1992). 18. J. A. Van Allen and B. A. Randall, J. Geophys. speed inside the terminal shock, V1) and the 1. E. N. Parker, Interplanetary Dynamical Processes fRes. 90,1399 (1985); W. R. Webber, J. A. Lock- (Interscience, New York, 1963). parameter 8 (the ratio of the distance to the wood, J. R. Jokipii, ibid. 91, 4103 (1986). 2. W. I. Axford, in Physics of the Outer Heliosphere, terminal shock, Rr, to the distance to the helio- 19. G. B. Nelson and D. B. Melrose, in Solar Radiophys- S. Grzedzielski and D. E. Page, Eds. (Pergamon, pause, RH). ics, D. J. McLean and N. R. Lebrum, Eds. (Cam- Oxford, 1990), 7-15. pp. bridge Univ. Press, Cambridge, 1985), pp. 333- 3. S. T. Suess, Rev. Geophys. 28, 97 (1990). 359. a = Vl 4. For a description of the Voyager PWS, see F. L. 8 , V21 20. W. M. Macek, I. H. Cairns, W. S. Kurth, D. A. Scarf and D. A. Gurnett, Space Sci. Rev. 21, 289 Res. (1991); 1. H. (Ri/H) (1977). Gurnett, Geophys. Lett. 18, 357 0.6 0.7 0.8 D. A. Res. 97, 5. J. Belcher, personal communication. Cairns and Gurnett, J. Geophys. 6235 (1992); I. H. Cairns, W. S. Kurth, D. A. 6. A. Czechowski and S. Grzedzielski, Nature 344, V1 = 800 km s-' Gurnett, ibid., p. 6245. 0.80 164 172 177 640 (1990). 7. W. S. Kurth, D. A. Gurnett, F. L. R. 21. V. B. Baranov, Space Sci. Rev. 52, 89 (1990); V. B. 0.75 160 168 Scarf, L. 175 Poynter, ibid. 312, 27 (1984). Baranov and Yu. G. Malama, J. Geophys. Res. 98, 0.70 155 165 173 15157 (1993); R. S. Steinolfson, V. J. Pizzo, T. 8. __ , Geophys. Res. Lett. 14, 49 (1987);W. S. Kurthand D. A. Gurnett, ibid. 18, 1801 (1991); W. Holzer, in preparation. V, = 600 km s-1 22. A. F. Davidsen, Science 259, 327 (1993); R. 0.80 123 129 133 S. Kurth, Adv. Space Res., in press. 9. R. L. McNutt, Geophys. Res. Lett. 15, 1307 Lallement, J.-L. Bertaux, J. T. Clark, ibid. 260, 0.75 120 126 131 (1988). 1095 (1993). 0.70 116 124 130 10. S. Grzedzielski and A. J. Lazarus, J. Geophys. 23. R. S. Steinolfson, personal communication. Res. 98, 5551 (1993). 24. J. M. Ajello, A. I. Stewart, G. E. Thomas, A. Garps, 11. J. D. Gosling et al., ibid. 79, 4581 (1974); S. W. Astrophys. J. 317, 964 (1987). Kahler, Annu. Rev. Astron. Astrophys. 30, 113 25. We thank the Voyager team at the Jet Propulsion 0.6 to 0.8. For the distance parameter 8, (1992); J. D. Gosling, J. Geophys. Res., in press. Laboratory (JPL) for their valuable support, espe- numerical simulations (21) suggest an av- 12. S. E. Forbush, Phys. Rev. 51,1108 (1937). cially E. Stone, E. Franzgrote, and S. Linick for 13. Solar-GeophysicalData, Prompt Reports, 562-Part 1 scheduling increased wideband coverage during erage value of about 0.75, with a range (National Oceanic and Atmospheric Administration, this event, and C. Avis and N. Toy for timely and from about 0.7 to 0.8. The calculations Boulder, CO, 1991); ibid., 563-Part 1. efficient production of the experiment data records. place the distance to the heliopause in the 14. J. A. Van Allen and R. W. Fillius, Geophys. Res. We also thank R. Lepping for his help in determining Lett. 19, 1423 (1992). the spacecraft attitude during the roll maneuvers, J. range from about 116 to 177 AU (Table 15. W. R. Webber and J. A. Lockwood, J. Geophys. A. Van Allen for numerous useful discussions, and L. 1). For the nominal value of 8, the dis- Res. 98, 7821 (1993); see also F. B. McDonald et Granroth, J. Cook-Granroth, and T. Barnett-Fisher tance to the terminal shock is about 87 to al., paper presented at the 23rd International for their efforts in processing the data. The research Cosmic Ray Conference, Calgary, Canada, 19 to at the University of Iowa was supported by the 133 AU. With the use of numerical sim- 30 July 1993. National Aeronautics and Space Administration ulations, it should be possible to greatly 16. L. F. Burlaga et al., J. Geophys. Res. 89, 6579 through contract 959193 with JPL.

SCIENCE * VOL. 262 * 8 OCTOBER 1993 203