Four years of SOHO Discoveries-Some Highlights - Nasa

ulletin 102 — may 2000 bull68

ulletin 102 — may 2000 bullTable 1. The SOHO scientific instrumentsInvestigationPrincipal InvestigatorGOLF Global Oscillations at Low Frequencies A. Gabriel, IAS, Orsay, FranceVIRGO Variability of Solar Irradiance and Gravity Oscillations C. Fröhlich, PMOD Davos, SwitzerlandMDI Michelson Doppler Imager P. Scherrer, Stanford University, USASUMER Solar Ultraviolet Measurements of Emitted Radiation K. Wilhelm, MPAe Lindau, GermanyCDS Coronal Diagnostic Spectrometer R. Harrison, RAL, Chilton, UKEIT Extreme-Ultraviolet Imaging Telescope J.-P. Delaboudinière, IAS, Orsay, FranceUVCS Ultra-Violet Coronagraph Spectrometer J. Kohl, SAO, Cambridge, USALASCO Large-Angle Spectroscopic Coronagraph R. Howard, NRL, Washington, USASWAN Solar Wind Anisotropies J.-L. Bertaux, SA, Verrières, FranceCELIAS Charge, Element and Isotope Analysis System P. Bochsler, Univ. of Bern, SwitzerlandCOSTEP Comprehensive Supra-Thermal and Energetic-Particle Analyser H. Kunow, Univ. of Kiel, GermanyERNE Energetic and Relativistic Nuclei and Electron Experiment J. Torsti, Univ. of Turku, FinlandIAS : Institut d’Astrophysique Spatiale SAO : Smithsonian Astrophysical ObservatoryPMOD : Physikalisch-Meteorologisches Observatorium Davos NRL : Naval Research LaboratoryRAL : Rutherford Appleton Laboratory SA : Service d’AeronomieMPAe: Max-Planck-Institut für Aeronomiecan infer the temperature, density, elementaland isotopic abundances, interior mixing,interior rotation and flows, even the age of theSolar System, and pursue such esotericmatters as testing the constancy of thegravitational constant.Interior rotation and flowsThe nearly uninterrupted data from SOHO’sMichelson Doppler Imager (MDI) yield oscillationpower spectra with an unprecedented signalto-noiseratio that allow the determination ofthe frequency splittings of the global resonantacoustic modes of the Sun with exceptionalaccuracy. These data confirm that the decreasein angular velocity Ω with latitude seen at thesurface extends with little radial variationthrough much of the convection zone, at thebase of which is an adjustment layer, called the‘tachocline’, leading to nearly uniform rotationdeeper in the radiative interior (Fig. 1).Furthermore, a prominent rotational shearinglayer in which Ω increases just below the surfaceis discernible at low- to mid-latitudes.The MDI team has also been able to study thesolar rotation closer to the poles than has beenachieved in previous investigations. The datahave revealed that the angular velocity isdistinctly lower at high latitudes than previouslyextrapolated from measurements at lowerlatitudes based on surface Doppler observationsand helioseismology. Moreover, they foundevidence of a submerged polar jet nearlatitudes of 75°, which is rotating more rapidlythan its immediate surroundings (red oval nearthe poles in Fig. 2).Alternating zonal bands of faster and slowerrotation (± 7 m/s) at a depth of 2 – 9 Mmappear to coincide with an evolving pattern of‘torsional oscillations’ reported from earliersurface Doppler studies (Fig. 2). Clear evidenceof the migration of these zonal flows towardsthe equator has been found, and a recentstudy has established that these banded flowsare not merely a near-surface phenomenon.Rather, they extend downward at least 60 Mm(some 8% of the total solar radius) and thusare evident over a significant fraction of thenearly 200 Mm depth of the solar convectionzone (Fig. 3).Long-lived velocity cells extending over 40–50°of longitude, but less than 10° of latitude, havebeen identified with the elusive ‘giant cells’ byBeck et al.. Their suprisingly large aspect ratiomay be a consequence of the Sun’s differentialrotation, whereby larger features are broken upby rotational shear.High-precision MDI measurements of the Sun’sshape obtained during two special 360° rollmanoeuvres of the SOHO spacecraft have70

ulletin 102 — may 2000 bullFigure 6. A vertical cutthrough the upperconvection zone showingsubsurface flows andsound speedinhomogeneities (fromKosovichev et al.)Figure 7. The geometryof the time-distanceanalysis of subsurfacemeridional flows (fromGiles et al.)persist to great depths (Fig. 7), and thereforemay play an important role in the 11-year solarcycle. They found the meridional flow to persistto a depth of at least 26 Mm, with a depthaveragedvelocity of 23.5 ± 0.6 m/s at midlatitude.More recently, they extended thesemeasurements to a depth of 0.8 R withoutfinding any evidence of a return flow. Continuityconsiderations led them to estimate the returnflow below 0.8 R at approximately 5 m/s,which might actually be detectable in thefuture, providing a useful constraint for dynamotheories.One of the most exciting applications of solartomography is in studying the birth andevolution of active regions and complexes ofsolar activity. Kosovichev et al. have studiedthe emergence of an active region on the Sun’sdisc with this technique and their resultssuggest that the emerging flux ropes travel veryquickly through the upper 18 Mm of theconvection zone. They estimate the speed ofemergence at about 1.3 km/s, which issomewhat higher than predicted by earliertheories. Wave speeds vary in the emergingactive region by about 0.5 km/s. The observeddevelopment of the active region suggests thatthe sunspots are formed as a result of theconcentration of magnetic flux close to thesurface. The Stanford team also presentedtime-distance results on the subsurfacestructure of a large sunspot observed on20 June 1998 (Fig. 8). The wave-speedperturbations in the spot are much stronger thanin the emerging flux (0.3–1 km/s). At a depth of4 Mm, a 1 km/s wave-speed perturbationcorresponds to a 10% temperature variation(approx. 2800 K) or to a 18 kG magnetic field.Beneath the spot, the perturbation is negativein the subsurface layers and becomes positivefurther down in the interior. Their tomographicimages also revealed sunspot ‘fingers’ – long,narrow structures at a depth of about 4 Mm –depth (Mm)horizontal distance (Mm)74

soho discoverieswhich connect the sunspot with surroundingpores of the same polarity. Pores with theopposite polarity are not connected to the spot.MDI has also made the first observations ofseismic waves from a solar flare, opening uppossibilities for studying both the flares and thesolar interior. During the impulsive phase of theX2.6-class flare of 9 July 1996, a high-energyelectron beam caused an explosive evaporationof chromospheric plasma at supersonicvelocities. The upward motion was balanced bya downward recoil in the lower chromosphere,which excited propagating waves in the solarinterior. On the surface, the outgoing circularflare waves resembled ripples from a pebblethrown into a pond (Fig. 9). The seismic wavepropagated at least 120 000 km from the flare’sepicentre, with an average speed of about50 km/s on the solar surface.Transition-region dynamicsExplosive events and ‘blinkers’Several types of transient events have beendetected in the quiet Sun. High-velocity eventsin the solar transition region, also called‘explosive events’, were first discovered in theearly eighties based on ultraviolet observationswith the High-Resolution Telescope-Spectrometer(HRTS) rocket payload. They have largevelocity dispersions, approximately ±100 km/s,i.e. velocities are directed both towards andaway from the observer causing a strongbroadening of the spectral lines observed.Explosive events have been studied extensivelyby a number of authors using SUMER data,and several results support the magneticreconnection origin of these features. Innes etal. have reported explosive events that showspatially separated blue- and red-shifted jetsand some that show transverse motion of blueand red shifts, as predicted if reconnectionwas the source (Fig. 10). Comparisons withmagnetograms from MDI and those obtainedat ground-based solar observatories have alsoprovided evidence that transition-region explosiveevents are a manifestation of magneticreconnection occurring in the quiet Sun.Harrison et al. have presented a comprehensivestudy of EUV flashes, also known as ‘blinkers’,Figure 8. The subsurface sound speedperturbations in a sunspot region observed on20 June 1998 by MDI. The horizontal size ofthe box is 13 deg (158 Mm), the depth is24 Mm. The horizontal cut in panels (a) and (b)reaches down to a depth of 21.6 Mm, while inpanel (c) it is 4.8 Mm deep. Positive variationsare shown in red, negative variations in blue.Scaling: ±1 km/s (from Kosovichev et al.)75

ulletin 102 — may 2000 bullFigure 9. Seismic waves(‘sun quake’) produced bya solar flare on 9 July1996 (from Kosovichev &Zharkova)y, Mmy, Mmx, Mm x, MmFigure 10. Bi-directionalplasma jets observed bySUMER in Si IV 1393 Å inJune 1996 and aschematic of the plasmaflow (from Innes et al.)which were identified in the quiet Sun network asintensity enhancements of order 10–40% usingCDS (Fig. 11). They have analysed 97 blinkerevents and identified blinker spectral, temporaland spatial characteristics, their distribution,frequency and general properties, across abroad range of temperatures, from 20 000 to1 200 000 K. The blinkers are most pronouncedin the transition region lines O III, O IV and O V,with modest or no detectable signature athigher and lower temperatures. A typical blinkerlasts about 1000 s, but due to a long tail oflonger duration events the average duration is2400 s. Comparisons with plasma coolingtimes led to the conclusion that there must becontinuous energy input throughout the blinkerevent. There are about 3000 blinker events inprogress at any given time. Remarkably, lineratios from O III, O IV and O V show nosignificant change throughout the blinker event,suggesting that the intensity increase is not atemperature effect, but is predominantlycaused by increases in density or filling factor.The thermal-energy content of an averageblinker is estimated as 2 x 10 25 erg.While the explosive events appear as extremelybroad line profiles with Doppler shifts of76

ulletin 102 — may 2000 bullMore recent investigations using SUMERobservations have revisited this problem,addressing possible errors in the restwavelengths of lines from highly ionised atoms(e.g. Ne VIII, Na IX, Mg X, Fe XII). Using full-discscans from SUMER and assuming that all massor wave-motion effects on the limb cancel outstatistically, new rest wavelengths for Ne VIIIand Mg X have been established, leading toblueshifts of 2.5 km/s and 4.5 km/s, respectively,at disc centre.These recent results suggest that the uppertransition region and lower corona appearblue-shifted in the quiet Sun, with a steeptransition from red- to blue-shifts above 5 x 10 5 K.This transition is significant because it hasmajor implications for the transition region andsolar-wind modelling, as well as for ourunderstanding of the structure of the solaratmosphere.The networkEarly models of the solar atmosphere assumedthat the temperature structure of the upperatmosphere was continuous, with a thintransition region connecting the chromospherewith the corona. This depiction now appearstoo simplistic. Rather, it seems that the solaratmosphere consists of a hierarchy ofisothermal, highly dynamic loop structures.Of particular interest in this context is thenetwork, which is believed to be the backboneof the entire solar atmosphere and the basicchannel of the energy responsible for heatingthe corona and accelerating the solar wind.Patsourakos et al. have used CDS data tostudy the width variation of the network withtemperature. They found that the networkboundaries have an almost constant width upto about 250 000 K (where the network contrastis also strongest) and then fan out rapidly atcoronal temperatures. The network in thelower transition region is about 10 arcsecacross and spreads to about 16 arcsec at1 MK. These results are in very good agreementwith Gabriel’s transition region-corona model,dating from1976.Active-region dynamicsEIT, SUMER and CDS observations haveclearly demonstrated that the solar transitionregion and corona are extremely dynamic andtime variable in nature. Large line shifts of up to60 km/s were observed with CDS in individualactive region loops (Fig. 13). High Dopplershifts are common in active-region loops andstrong shifts are present in parts of loops fortemperatures up to 0.5 MK. Regions with bothred and blue shifts are seen. While typicalvalues correspond to velocities of ± 50-100 km/s,shifts approaching 200 km/s have beendetected. At temperatures T > 1 MK, i.e. inMg IX 368 Å or Fe XVI 360 Å, only small shiftsare seen. The high Doppler shifts thereforeseem to be restricted to the transition region.Brynildsen et al. studied 3-min transition-regionoscillations above sunspots by analysing timeseries recorded in O V 629 Å, N V 1238 Å and1242 Å, and the chromospheric Si II 1260 Åline in NOAA 8378. The 3-min oscillations thatFigure 13. CoronalDiagnostic Spectrometer(CDS) raster imagesshowing the intensity andvelocity distribution oftransition-region plasmain AR8737 on the southwestlimb in October1999. Six emission lineswere observedsimultaneously, of whichone is shown here(O V 629 Å, formed atabout 230 000 K). Dopplershifts corresponding to avelocity greater than± 40 km/s (144 000 km/h)are fully red/blue.Contours outline areaswith velocities exceeding± 50 km/s (180 000 km/h)(courtesy of T. Fredvik &O. Kjeldseth-Moe)78

soho discoveriesFigure 16. Magneticcarpet. Model of magneticfield lines based on MDImagnetograms,superimposed on animage of the solar coronain Fe XII 195 Å from EIT(courtesy of SOHO/MDIConsortium)distribution of apparent locations of CMEsvaried during this period as expected. While thepointing stability provided by the SOHOplatform in its L-1 orbit and the use of CCDdetectors have resulted in superior brightnesssensitivity for LASCO over earlier coronagraphs,they have not detected a significant populationof fainter (i.e. low-mass) CMEs. The generalshape of the distribution of apparent sizes forLASCO CMEs is similar to those of earlierreports, but the average (median) apparent sizeof 72° (vs. 50°) is significantly larger.St.Cyr et al. have also reported on a populationof CMEs with large apparent sizes, whichappear to have a significant longitudinalcomponent directed along the Sun–Earth line,either toward or away from the Earth. These arethe so-called ‘halo CMEs’ (Fig. 18). Using fulldiscEIT images, they found that 40 out of 92 ofthese events might have been directed towardsthe Earth. A comparison of the timing of thoseevents with the Kp geomagnetic storm index inthe days following the CME showed that 15out of 21 (71%) of the Kp > 6 storms could beaccounted for as SOHO LASCO/EIT front-sidehalo CMEs. Three more Kp storms may haveFigure 17. Progress of aCoronal Mass Ejection(CME) observed over an8 h period on 5-6 August1999 by LASCO C3. Thedark disc blocks the Sunso that the LASCOinstrument can observethe structures of thecorona in visible light.The white circlerepresents the size andposition of the Sun(courtesy of SOHO/LASCOConsortium)81

ulletin 102 — may 2000 bullFigure 18. Massive ‘halo’CME as recorded byLASCO C2 on 17 February2000. Displayed here is aso-called ‘runningdifference’ image,showing the variation inbrightness from oneframe to the next(courtesy of SOHO/LASCOConsortium)Figure 19. Sequence ofEIT difference imagesshowing the intensity(density) enhancementand following rarefactionassociated with a shockwave expanding acrossthe solar disc from thesite of the origin of aCME, recorded on 12 May1997. A halo CME wasobserved by LASCO.These images wereformed from thedifferences of successiveimages in the emissionlines of Fe XII near 195 Å;this ion is formed attemperatures of about 1.5million degrees. The wavefront travels at speeds of~ 300 km/s, typical of afast mode Alfvén shock inthe lower solar corona(from Thompson et al.)been missed during LASCO/EIT datagaps, bringing the possible associationrate to 18 out of 21 (86%).EIT has discovered large-scale transientwaves in the corona, also called‘Coronal Moreton Waves’ or ‘EITwaves’, propagating outward fromactive regions below CMEs. Theseevents are usually recorded in theFe XII 195 Å bandpass, during highcadence(< 20 min) observations. Theirappearance is stunning in that theyusually affect most of the visible solardisc (Fig. 19). They generally propagateat speeds of 200–500 km/s, traversinga solar diameter in less than an hour.Active regions distort the waves locally,bending them towards the lower Alfvénspeed regions. On the basis of speedand propagation characteristics, the EIT waveswere associated with fast-mode shock waves.Another interesting aspect of these coronalMoreton waves is their association with theacceleration and injection of high-energyelectrons and protons, as measured, forexample, by COSTEP and ERNE.Solar windOrigin and speed profile of the fast windCoronal-hole outflow velocity maps obtainedwith the SUMER instrument in the Ne VIIIemission line at 770 Å show a clear relationshipbetween coronal-hole outflow velocity and thechromospheric network structure (Fig. 20), withthe largest outflow velocities occurring alongnetwork boundaries and at the intersection ofnetwork boundaries. This can be consideredthe first direct spectroscopic determination ofthe source regions of the fast solar wind incoronal holes.Proton and O VI outflow velocities in coronalholes have been measured by UVCS using theDoppler dimming method. The O VI outflowvelocity was found to be significantly higherthan the proton velocity, with a very steepincrease between 1.5 and 2.5 R , reachingoutflow velocities of 300 km/s at around 2 R (Fig. 21). While the hydrogen outflow velocitiesare still consistent with some conventionaltheoretical models for polar wind acceleration,the higher oxygen flow speeds cannot beexplained by these models. A possibleexplanation is offered by the dissipation of highfrequencyAlfvén waves via gyroresonance withion-cyclotron Larmor motions, which can heatand accelerate ions differently depending ontheir charge and mass.Speed profile of the slow solar windTime-lapse sequences of LASCO white-lightcoronagraph images give the impression of acontinuous outflow of material in the streamerbelt. Density enhancements, or ‘blobs’, formnear the cusps of helmet streamers and appearto be carried outward by the ambient solarwind. Sheeley et al., using data from theLASCO C2 and C3 coronagraphs, have traceda large number of such ‘blobs’ from 2 to over25 solar radii. Assuming that these ‘blobs’ arecarried away by the solar wind like leaves on82

soho discoveriesthe river, they have measured the accelerationprofile of the slow solar wind, which typicallydoubles its speed from 150 km/s near 5 R to300 km/s near 25 R . They found a constantacceleration of about 4 ms -2 through most ofthe 30 R field-of-view. The speed profile isconsistent with an iso-thermal solar-windexpansion at a temperature of about 1.1 MKand a sonic point near 5 R .Solar-wind compositionUsing data from the CELIAS/MTOF sensor, theCELIAS team has made the first in-situdetermination of the isotopic composition ofcalcium and nitrogen in the solar wind. Thesemeasurements are important for studies ofstellar modelling and Solar System formation,because the present-day solar Ca isotopicabundances are unchanged from their originalisotopic composition in the solar nebula. Theisotopic ratios 40Ca/ 42 Ca and 40Ca/ 44 Cameasured in the solar wind were found to beconsistent with terrestrial values. The isotoperatio 14 N/ 15 N was found to be 200 ± 60,indicating a depletion of 15 N in the terrestrialatmosphere compared to solar matter.Ion freeze-in temperatures were measured byCELIAS/CTOF with a time resolution of 5 min.These measurements indicate that some of thefilamentary structures of the inner coronaobserved in H α survive in the inter-planetarymedium as far as 1 AU.Figure 20. Source regions of the fast solar wind. Background: EIT full-Sun imagetaken in the emission line of Fe XII 195 Å, revealing gas at 1.5 million degreesshaped by magnetic fields. Bright regions indicate hot, dense plasma loopswith strong magnetic fields, while dark regions imply an open magnetic fieldgeometry, and are the source of the high-speed solar wind. The ‘zoomed-in’or ‘close-up’ region shows a Doppler velocity map of plasma at about630 000 K at the base of the corona, as recorded by SUMER in the Ne VIIIemission line at 770 Å. Blue represents blue shifts or outflows and redrepresents red shifts or downflows. The blue regions are inside a coronal hole,or open magnetic field region, where the high-speed solar wind is accelerated.Superposed are the edges of ‘honeycomb-shaped’ patterns of magnetic fieldsat the surface of the Sun (from Hassler et al.)The unprecedented time resolution of theCELIAS/CTOF data has allowed a fine-scaledstudy of the elemental Fe/O ratio as a functionof the solar-wind bulk speed. Since Fe is a lowFirst Ionisation Potential (FIP) element and O ahigh-FIP element, their relative abundance isdiagnostic for the so-called ‘FIP fractionationprocess’. The Fe/O abundance shows acontinuos decrease with increasing solar-windspeed by a factor of two between 350 km/sand 500 km/s, in correspondence with thewell-established FIP effect.CometsSOHO is not only providing new measurementsabout the Sun. On 4 February 2000, SOHOdiscovered its 100th comet, 93 of which belongto the Sun-grazing Kreutz family (Fig. 22). Aparticular feature is the presence of a dust tailfor only a few Sun grazers. Analysis of the lightcurves is used to investigate the properties ofthe nuclei (size, fragmentation, destruction) andthe dust production rates.Thanks to rapid communication from theLASCO group and the near-real-time observingcapabilities of the SOHO instruments, UVCScould make spectroscopy measurements ofFigure 21. Empirical outflow velocity of O VI and H I in coronal holes over thepoles (from Kohl et al.)83

ulletin 102 — may 2000 bullseveral comets on the day of their discovery.UVCS measurements of comet C/1996Y1obtained at 6.8 R confirmed the predictions ofmodels of the cometary bow shock driven bymass-loading as cometary molecules areionised and swept up in the solar wind. Fromthe width and shift of the line profiles, the solarwindspeed at 6.8 R could be determined (640km/s). The outgassing rate of the comet wasestimated at 20 kg/s, implying an activenucleus area of only about 6.7 m in diameterand a mass of about 120 000 kg.Figure 22. LASCO sees two comets plunge into the Sun. In a rare celestialspectacle, two comets were observed by the LASCO coronagraph plunginginto the Sun’s atmosphere in close succession, on 1 and 2 June 1998.Science instruments on SOHO have discovered more than 100 comets,including many so-called ‘Sun grazers’, but none in such close succession(courtesy of SOHO/LASCO Consortium)Figure 23. SWAN H I Ly-α image of the huge cloud of hydrogen surroundingcomet Hale-Bopp when it neared the Sun in the spring of 1997. The smallyellow dot shows the Sun to scale (from Combi et al.)Comets are surrounded by large clouds ofhydrogen, produced by the break-up of watermolecules evaporating from the comets’ ice.The solar-wind mapper SWAN sees these largeclouds of hydrogen glowing in the light of theH I Lyman-α line. The huge cloud of hydrogensurrounding Comet Hale-Bopp (Fig. 23) duringits perihelion passage in the spring of 1997 wasmore than 100 million kilometres wide,diminishing in intensity outwards (contour lines).It far exceeded the great comet’s visible tail(inset photograph). Although generated by acomet nucleus perhaps only 40 km in diameter,the hydrogen cloud was 70 times wider thanthe Sun itself (yellow dot to scale) and ten timeswider than the hydrogen cloud of CometHyakutake observed by SWAN in 1996. Thewater evaporation rate of Hale-Bopp wasmeasured by SWAN at more than 200 milliontons per day. Comet Wirtanen, the target forESA’s Rosetta mission (2003), pumped outwater vapour at a rate of 20 000 tons a dayduring its most recent periodic visit to the Sun,according to the SWAN data. SWAN has alsoseen something else extraordinary – thebiggest shadow ever observed in our SolarSystem, namely that of a comet projected onthe sky behind it (Fig. 24).HeliosphereThe Sun is moving through the Local InterstellarCloud (LIC) at about 26 km/s. The solar windbuilds a cavity, the heliosphere, within theionised gas component of the LIC. The neutralatoms (e.g. He) of the LIC, on the other hand,enter the heliosphere unaffected. The He flowproperties are now well-constrained from aseries of measurements (v He= 25.5 ± 0.5 km/s,T He= 6000 ± 1000 K). These values are inagreement with the LIC velocity andtemperature deduced from stellar spectroscopy.Hydrogen, on the other hand, isexpected to be affected by coupling with thedecelerated plasma via charge-exchange.Neutral hydrogen heating and decelerationtherefore provides a measurement of thiscoupling and, in turn, of the plasma density inthe LIC, which is responsible for most of theheliosphere’s confinement.84

soho discoveriesThe SWAN team, analysing data fromabsorption cells, found hydrogen temperaturesT 0of 11 500 ± 1500 K, i.e. significantly abovethe temperature of the interstellar He flow (6000± 1000 K), requiring strong heating of morethan 3500 K at the heliosphere interface. Partof this excess temperature is probably due toradiative-transfer effects.The apparent interstellar hydrogen velocity inthe up- and downwind direction was measuredto be -25.4 ±1 km/s and +21.6 ±1.3 km/s,respectively, with the most precise determination(since model-independent) of the H flowdirection. The new estimate of the upwinddirection from SWAN measurements is252.3 ± 0.73 deg and 8.7 ± 0.90 deg in eclipticcoordinates, which is off by about 3–4 deg fromthe He flow direction. The SWAN teamspeculates that this might be a sign of anasymmetry in the heliospheric interface due tothe ambient interstellar magnetic field.Comparing the above hydrogen temperatureand velocity measurements by SWAN withheliospheric models leads to an estimate of theinterstellar plasma density of n e ≈ 0.04 cm -3 . Itis interesting to note that the plasma frequencyfor n e ≈ 0.04 cm -3 is 1.8 kHz, i.e. exactly thevalue of the remarkably stable cut-off frequencyobserved by Voyager.Of particular interest for future studies might bethe temperature minimum measured betweenthe upwind and downwind directions. Classicalmodels predict a monotonic increase in theline-of-sight temperature from upwind todownwind. The authors interpret this behaviouras first evidence of the existence of two distinctpopulations at different velocities, as predictedby some heliosphere/interstellar-gas interfacemodels. If confirmed, this should provide agood diagnostic of the interface.Total solar irradiance variationsThe VIRGO instrument on SOHO extends therecord of Total Solar Irradiance (TSI)measurements into cycle 23. In Figure 25,measurements from six independent spacebasedradiometers since 1978 (top) have beencombined to produce the composite TSI overtwo decades (bottom). They show that theSun’s output fluctuates during each 11-yearsunspot cycle, changing by about 0.1%between maxima (1980 and 1990) and minima(1987 and 1997) in solar activity. Temporarydips of up to 0.3% and a few days duration arethe result of large sunspots passing over theFigure 24. SWANobservations of cometHale-Bopp’s shadow(courtesy of SOHO/SWANConsortium)85

ulletin 102 — may 2000 bullFigure 25. Total solarirradiance variations from1978 to 1999. The dataare from the Hickey-Frieden (HF) radiometer ofthe Earth RadiationBudget (ERB) experimenton Nimbus-7 (1978-1992),the two Active CavityRadiometer IrradianceMonitors (ACRIM I and II)aboard the SolarMaximum Mission (1980-1989) and the UpperAtmosphere ResearchSatellite (1991-),respectively, and theVIRGO radiometers onSOHO (1996-). Also shownare the data from theradiometer on the ERB(1984-), and SOVA2 aspart of the SolarVariability Experiment onthe European RetrievableCarrier (1992-1993) (fromQuinn and Fröhlich)visible hemisphere. The larger number ofsunspots near the peak in the 11-year cycle isaccompanied by a general rise in magneticactivity that creates an increase in the luminousoutput which exceeds the cooling effects ofsunspots. Offsets between the various datasets are the direct result of uncertainties in theabsolute radiometer scale of the radiometers(±0.3%). Despite these biases, each data setclearly shows varying radiation levels that trackthe overall 11-year solar activity cycle.ConclusionsSOHO set out to tackle three broad topics insolar and heliospheric physics: the structureand dynamics of the solar interior, the heatingand dynamics of the solar corona, and theacceleration and composition of the solar wind.In all three areas, its observations have allowedgreat strides to be made in our understandingof the diverse physical processes at work inour Sun. This has been made possible bythe comprehensive suite of state-of-the-artinstruments mounted on the superb and stableSOHO spacecraft, operating from the uniquevantage point of the L1 halo orbit.In such complex areas of research as solarphysics, progress is not made by just a fewpeople acting in a vacuum. The scientificachievements of the SOHO mission are theresults of a concerted, multi-disciplinary effortby a large international community of solarscientists, involving sound investments inspace hardware, coupled with a vigorous andwell-coordinated scientific operation andinterpretation effort. The interplay betweentheory and observations has already providedmany new insights and will continue to do sofor many years.With the wealth of SOHO data already in thearchive (and many more data yet to come,hopefully well beyond the next solar maximum),we should be able to unravel even more of themysteries of our closest star.AcknowledgementsThe great success of the SOHO mission is atribute to the many people who designed andbuilt this exquisite spacecraft and theseexcellent instruments, and to the many peoplewho diligently work behind the scenes to keepit up and running. Special thanks go to: HaroldW. Benefield and his AlliedSignal FlightOperations Team; Helmut Schweitzer andJean-Philippe Olive from the ESA/MMS TechnicalSupport Team; the Science Operations CoordinatorsLaura Roberts, Joan Hollis and PietMartens; Craig Roberts, John Rowe and theircolleagues from Flight Dynamics, the colleaguesfrom DSN, and, last but not least, to FrancisVandenbussche and his recovery team formaking a miracle come true! r86