zotero-db/storage/TAFK9IZ6/.zotero-ft-cache

GEOPHYSICALRESEARCHLETTERS,VOL. 25,NOo20, PAGES3787-3790,OCTOBER15, 1998

Effects of a Mid-Latitude Solar Eclipse on the
Thermosphere and Ionosphere- A Modelling Study
I.C.F. Miiller-Wodarg,A.D. Aylward
AtmospherPichysicLsaboratorDy,epartmenotfPhysicasndAstronomUyn, iversitCyollegLeondon
M. Lockwood SpaceSciencDe epartmentR, utherfordAppletonLaboratory

Abstract. A modellingstudy is presentedwhich inves- The CTIP Model

tigatesin-situgeneratecdhangeosfthethermosphearned The Coupled Thermosphere-IonosphereP-lasmasphere

ionospherdeuringa solareclipseN. eutraltemperatureasre expectedto dropby up to 40ø K at 240km heightin the totalityfootprintw, ith neutrawl indsof up to 26m/s respondintgo thechangoefpressureB.othtemperatureasnd windsarefoundto respondwith a time lagof 30 min after
the passingof the Moon'sshadow.A gravitywaveis generated in the neutral atmosphereand propagatesinto the
oppositheemisphearet around300m/s. Thecombineedf-

Model(CTIP) [Fuller-Rowelelt al., 1996;Millwardet al., 1996, and referencestherein]solvesself-consistentltyhe time-dependent,3-D coupledequationsof momentum,energyandcontinuityfor neutralparticlesand ions.Thermosphericcalculationsarecarriedoutona sphericacl o-rotating pressurceoordinatgeridwithspacingosf2ø in latitude,18ø in longitudeandonescaleheightvertically,between80 km and around 450 km altitude, at time steps of 60 seconds.

fectsof thermalcoolinganddownwellinlgeadto an overall The ionosphereusesa sphericalheight coordinatesystem

increasien [O],while[N2]initiallyrisesandthenfor sev- with the same horizontal resolution, but height levels rang-

eralhoursaftertheeclipseisbelowthe "steadystate"level. An enhancemeonft [NmF2]is foundandexplainedby the

ing from 100 to around10000km. To considerthe moon's shadow,the solar heating func-

atmospherec'osntractiodnuringa, ndthereduced[O]/[N2] tion wassetto zero in a footprint moving at the appropriate

ratio after the eclipse.

velocity(on average1.5 km/s) alongthe path of totality.

Introduction

The largeregionofhalf-shadowwasalsoconsidereda,ssuminganincreasoefsolarluminositbyy3 %ofthenon-eclipsed

The followingstudy investigatetshe effectsof a so- valueper 100km distancefromthe totality footprint.The lar eclipseon the upperatmosphereu,singthe Coupled simulation was carried out for August 11, 1999, assuming ThermospheIroen- osphere-PlasmaspMhoedrel(CTIP).We a solarF10.7 flux indexof 190 and magneticactivity index investigateheforthcominsgolareclipsoefAugust11,1999, of Kp=2+. In orderto identifythe effectsof the eclipse, witha pathoftotalitystartingat 9:32UniversaTlime(UT) a further simulation was run for identical conditions, but near41.0ø N/65.1øW, southof NovaScotia.Themoon's excludingthe eclipseshadow.

shadowwill move acrossthe south- westerncorner of the

U.K. mainland at 10:10 UT and continue acrosscentral Results and Discussion

France(10:20UT), southerGnerman(y10:30UT), Austria

(10:4U0 T),Hungar(y10:5U0 T),Romani(a11:00UT),central Turkey(11:20UT) anddisappeanreartheBayofBen-
gal(20.6ø N, 77.5øE) at 12:34UT.

Resukspresentedin thefollowingaredifferencebsetween the eclipseandnon-eclipsseimulationw, ith positive(negative)valuesdenotinganincreas(edecreaseo)ftheparameter

Previousstudiesofeffectsin theupperatmospherecaused in the eclipsesimulation.

bysolareclipseinscludtehosebySinghet al., [1989a]nd Chengetal., [1992]F. rittsandLuo[1993i]nvestigatethde Thermospheric changes

generatioonfgravitywavesin themiddleatmosphedreur- Figurei showcshangeosfneutratlemperatur(esolidline)

ingsolareclipsesT.hesestudiedsiscusosbservatioenxsclu- and horizontalwinds (dashed:northward;dotted: east-

sivelyontheassumptiotnhat a solareclipsereduceoszone ward)at 240kmaltitudeforlatitudes30øN, 50øN and70ø

heatingin the middleatmosphereIt. isthusof interesto N, correspondintog locationsouthof, at andnorthof the

investigahteowfara reductioonfin-situthermosphearincd totality footprintat that particularUT (11:30):the peak

ionosphersicolarabsorptiomnightcontributteo eclipsef- temperaturedecreasiesof around30øK, 45øK and20øK,

fectsin theupperatmosphereR.obleet al., [1986c] arried respectivelya,t theselatitudes,correspondintog around2.5

outanequivalenstimulatiowniththeNCARThermosphere%, 3.8 % and 1.7 % of the averagebackgroundtemperature

GeneraCl irculationModel(TGCM), but didnot includea (1200øK). Thereductionin temperaturecausesa decrease

fully coupledionosphere.

of pressuroeverthe totalityfootprintto whichthe neutral

windsrespond.At 30ø N, the eclipsegeneratesnorthward

Copyrig1h9t98bytheAmericaGneophysiUcnailon.

windsreaching25 m/s, whileat 70øNthey aresouthward, reaching15m/s. Zonalwindchangeasrebelow10 m/s at

Papenr umbeGr RL-1998900045.
0094-8276/98/GRL-1998900045505.00

30øN andreach20 m/s at 70ø N. Considerintghat typical windspeedsfoundby the CTIP fortheselatitudesareup to

3787

19448007, 1998, 20, Downloaded from https://agupubs.onlinelibrary.wiley.com/doi/10.1029/1998GL900045, Wiley Online Library on [09/05/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License

3788

MOLLER-WODARGET. AL: EFFECTSOF SOLARECLIPSE

CTIP

Tn ondwindchonges

UT=11-30 240 km

However,the centreof the moon'sshadowcrossesthis longi-

tude at around 10:20 UT, soit is evident that there is a time

50 N

lag of around 30 minutesbetweenthe totality and the peak

60•''''

'''''

'.'' '' ' ' ' 30

temperature decrease; winds respond with the same time

40[E

,' ,V•

20 •

•..

"

0 __..,.._....-.;."c'"'--.-...v•,•_ _.

-20

'-

-40

20
10 •
0 •(D
E -10 •--•
-20

-60 ,, i,, i,, i,, i,, •

-30

-180-120-60

0 60 120 180

Longitude

lag. A further interestingfeature of Figure 2 is the global propagationof the temperature disturbance. One seesthat the changeof temperatureis strongestat 50ø N, wherethe path of totality crossesthe longitude0ø meridian. In addition, two wavefrontsare formed, one propagating over the north pole to the oppositelongitude(not shown)and the other equatorward into the southern hemisphere. The wave velocityis of around300 m/s and agreeswell with values measuredby $ingh et al., [1989]. Althoughmagnitudesof most temperature changesare not measureablewith any of

60 '' •'' 40

50 N • ' ' • ' ' • ' ' • ' ' 30
20

today's instruments it must be emphasizedthat the values presentedhere are a result of reduced in-situ absorption in the upper atmosphereonly. This issueis discussedfurther
in section 4.

20 ............\. ..

0 "': ....

:' '

-20

0 •
E -10•

-40 -60 ,, •,,
-180-120-60

•,, •,, •,, 0 60
Longitude

-20 • , , -30 120 180

Ionospheric and Neutral Composition changes
Changesoccurringnear the F2 peak altitudes at latitude 50ø N and longitude0ø are shownin Figure 3. The graphs are the changesof the F2 peak electrondensity([NmF2], solidline) and of N2 and O densities(dashedand dotted lines,respectively)on the pressurelevel closestto the F2 peak height. The beginning("first contact" at 9:50 UT)

70 N

60 ' ' • ' ' • ' ' • ' ' • ' ' • ' ' 30

40

20

and end ("fourth contact"at 11:16UT) of the eclipseare marked by vertical bars. The followingwill discussthe neutral compositionchangesand outline how these may affect the electrondensities. Although Figure 3 showsone partic-

20• .•.....................

10 •

0 :'• '',

"'. ' ':'•'• 0 •

. '
-20

. ••'' , •• -,•• ..""'

--10•E

-40

- '•'"'

- 20

ular location only, the behaviour was found to be similar for others along the path of totality.
One seesfrom Figure 3 that N2 densitiesincreaseduring the firsthalfof the eclipseby up to around3 ß10x3m-3, or around 5 %, then fall again and becomesmaller than the

-60 ,, •,, •,, •,, •,, • , , -30

"steady state" values for around 10 hours after the eclipse,

-180-120-60

0 60 120 180
Longitude

by up to 2. 1013m-3, or 3 %. The behaviourof the O
densitiesis differentin that they stay largerthan the "steady state" valuesthroughout the eclipseand for around 14 hours

Figure 1. The changesof neutral temperatureand winds
at 11:30 UT for 240 km altitude and 3 different latitudes, as predicted by the CTIP for the August 11, 1999 solar eclipse. Solid lines denote temperature, dashed and dotted lines northward- and eastward wind components,respec-
tively.

afterwards.Their largestincreaseoccursat around11:30 UT, reachingaround7- 1013m-3 , or 5 % ofthebackground
value.
As a result of the falling temperature, one would expect densitieson a level of fixed pressureto rise. It was shown earlier (seeFigure 2) that the peak temperaturedecrease occurs around 30 minutes after totality, or 10:50 UT. Al-

though it was there shownfor a fixed altitude, the temper-

ature behaviour on a fixed pressurelevel was found to be

around100 m/s, theseeclipsed-inducecdhangesin neutral very similar. If cooling were the only processinfluencing

winds are important. When producing a global snapshot densitiest,he curvesof [O] and [N•] in Figure3 shouldfall

of wind changes(not shown)oneseesan anticlockwisfelow after 10:50 UT, due to the temperature rise after totality.

around the low pressureregion. Ion drag plays a minor role However, the N• density already begins to fall before that,

sincethe footprint of totality is far from the magneticpole at 10:30and [O] beginsto fall only at 11:30UT. Socooling

and auroral oval, and the flow is thus to a reasonableapprox- alone cannot account for the observed behaviour.

imation geostrophic. The patterns of neutral temperature Another important processis the downwelling,or down-

and wind changesagreewell with those found by Roble et ward transport of gases.Downwellingis a result of the con-

al., [1986]for the May 30, 1984eclipse.Any discrepanciesverginghorizontal windswhich surroundthe eclipsedregion

can be explainedthrough the differencesin the paths, with and causea downward flux relative to pressurelevelsin or-

the 1984 eclipseoccurringat lowerlatitudes,giving a weaker derto conservemassIRishbethet al., 1969]. Coolingitself,

Coriolis force acting on the neutral winds.

without the downwelling,would causeno real composition

Figure 2 showsthe changeof temperature at longitude changeon a fixed pressurelevel, so the relative concentra-

0ø as a function of UT and latitude. The plot showsthat tionsof gascomponentws ouldremainunchangedIRishbeth

the peak temperature decreaseoccursat around 10:50 UT. andMiiller-Wodarg,1999].The downwellingh, oweverd, oes

19448007, 1998, 20, Downloaded from https://agupubs.onlinelibrary.wiley.com/doi/10.1029/1998GL900045, Wiley Online Library on [09/05/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License

MOLLER-WODARG ET. AL: EFFECTS OF SOLAR ECLIPSE

3789

CTIP

1999 Solar Eclipse 2Lo4n0gk=0mE value, correspondingto a change in the F2 layer critical

frequency,foF2, by around 0.2 MHz. These changesare

Tn changes

very small and hardly measurable, but neverthelessit is of

interestto discusswhat causesthem. Someof this [NmF2]

behaviour is due to the atmosphere's contraction and some

can be explained from the neutral composition changes. It

is well known that daytime F2 region electron densities are

sensitiveto neutral composition: production dependson the

O concentration, while the lossrate dependson mainly the

N2 concentration.As a rule of thumb, daytime [NmF2]

increaseswith the [O]/[N2] ratio. During the eclipse,no

photoionizationof O occurs,so the O density is irrelevant.

It was found earlier that N2 densities increase during the

eclipse,which shouldlead to a decreaseof [NmF2]. As

-30

with the neutral densities, though, the atmosphere's com-

pression(dueto cooling)alsocausesan increasein electron

-60

densities on a fixed plasma pressure level. So, the two pro-

cesses("thermalincrease"and "chemicalreduction")com-

-90

,

9 10 11 12 13 14 15 16 17 18 Universal Time

pete against each other. The fastestreaction leading to recombinationof ionsand
electrons is via N2 and NO+, and reaction rate calculation
revealthat an increasein IN2]by 5 % wouldover30 minutes

Figure

2.

Temperaturechangesduringthe August11,

leadto a reductionof [NmF2]byaround1%. So,the"chem-
ical reduction" is too weak to cause an overall decrease of

1999solareclipseat longitude0ø for 240 km altitude.

IN2]in Figure3.

Initially, after the temperature minimum at 10:50 UT,

changethe compositionsinceair from higherpressurelevelswith differentrelative gasconcentrationsis transported downwards. One seesin Figure 3 that densitiesof O and

[NmF2]doesfall againasa resultof the thermalexpansion. About an hour later, though, photoionizationunder the enhanced[O]/[N2]ratiocausesan overallincreasein [NmF2]

N2 increaseduringthe first half of the eclipse,between9:50 UT and around 10:20 UT. As shownabove, thesedensity in-

to its secondmaximum, at around 13:30. The slow decrease
of [O]/[N2] (dueto moleculardiffusion)duringthe after-

creasesare consistentwith the cooling. When plotting the noonhoursthen causesthe observedreductionof [NmF2].

changeof the [O]/[N2]ratio (not shown)onewill however
seethat the relative concentration of O increases,which is

In summary,the initial increaseof [NmF2] is a result of the atmosphere'scompressionw, hilethe behaviourafter the

what onewouldexpectfromthe downwellingc,arrying[O]richgasesfromhigherto lowerpressurelevels.This shows that both processesc,oolingand downwellingp, lay a role here. The downwellingwill upsetdiffusivebalance,somolec-
ular diffusionwill act to restorethe equilibrium. Molecular

eclipseis linked to the neutral composition. Similar to the earlier findings about temperature pertur-
bations, the electrondensity changespropagate as a wave awayfromthe path of totality (not shownhere). However, disturbancespropagateinto the equatorial regionsonly and

diffusionis thus the third processdetermining neutral gas

densitiesin Figure 3. When discussingthe IN2] behaviour,it is importantto CTIP

F2 layer changes

50N/0E

notethat coolinganddownwellinghavecompetingeffectson

its concentrationw: hilecoolingleadsto an increaseof [N2], downwellindgecreaseist. As a resultof the cooling,IN2]

Ix1_0.1_4.....•4.0x1010

rises between 9:50 and 10:20 UT. After that, downwelling

hasbecomestrongenoughto dominateeffectsof the cooling

and the N2 concentrationbeginsto drop. As temperature

increasesagain, after 10:50UT, the atmosphereexpands

? and further reduces the densities. As a result of both the
downwellingand expansion,IN2] dropsbelowthe steady
state level after 11:20 UT. Molecular diffusion of N2 into

•"\\

0

\\

11'0•'x• 10 0

the previouslyeclipsedregioneventuallyleadsto an overall riseof IN2]after 14:30.

IN2] "_ ...- .......

-'

_1.0x1010

The combinedeffectsof downwellingand coolingleadto a
sharpriseof [O]after9:50UT. Afterthe temperatureminimumat 10:50UT, the O densityin Figure3 stillincreasefsor

,, I1,,,II,, I,,, I,I, II,, I,I, II,, I,, -2.0xT1M0

9

10

11

12

13

14

15

16

17

18

UT

sometime (dueto the downwelling)b,ut at a reducedrate

(dueto the heating),maximizingat 11:30andthenfalling. Figure 3. Changesof parametersat and nearthe F2 peak

The continuoudsecreasoef [O]after 11:30is a resultof the height during the August 11, 1999 solar eclipseat latitude

atmosphere'sexpansionand moleculardiffusion.

50øN andlongitude0ø. The panelshows[NmF2](solid)as

In Figure3, [NmF2](solidline) increasesduringthe wellasO (dotted)andN2 densities(dashed)at the pressure eclipseby up to 4. 10•øm-s, or 8 % of the backgroundlevecl losestot theF2peak.All densitieasregivenin [m-3].

19448007, 1998, 20, Downloaded from https://agupubs.onlinelibrary.wiley.com/doi/10.1029/1998GL900045, Wiley Online Library on [09/05/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License

3790

MOLLER-WODARGET. AL:EFFECTSOF SOLARECLIPSE

not acrossthe north pole. The high-latitude electric field and auroral oval destroyany eclipse-drivenionosphericdis-
turbances.
Conclusions and possible limitations

ing the viability of observationisn the upper atmosphere duringthe event. The predictionsand modellingare the mostself-consistenytet madeand any differencetso the observationws ill providevaluableinformationp, articularlyon couplingto loweratmosphericlayers.

The abovefindingsprovide a morphologicalinsightinto processesoccurring in the thermosphereand ionosphereas a result of a temporary reductionin the flux of heatingand ionizing radiation. It is of interest to discusswhere possible limitations of the simulations may lie in relation to solar eclipses.
As mentioned previously,the presentedsimulationsdo

References
Cheng,K., Y.-N. Huang,and S.-W. Chen,IonosphericEffectsof theSolarEclipseofSeptembe2r3, 1987,AroundtheEquatorial AnomalyCrestRegion,J. GeophysR. es.,97, 103-111,1992.
Fritts, D.C., and Z. Luo, Gravity Wave Forcingin the Middle AtmosphereDue to ReducedOzoneHeatingDuringa Solar Eclipse,J. GeophysR. es., 98, 3011-3021,1993.

not considerany eclipsecontribution from the middle atmo- Fuller-Rowell,T.J., D. Rees, S. Quegan, R.J. Moffett, M. V.

sphere, where ozone absorption is strong, so effectsassociated with that will not be seenhere. This has advantagesin that one can clearly separate the contributions from below and in-situ. Overall, the only measurablechangesfound

Codrescu,and G. H. Millward, A CoupledThermosphereIonosphereModel (CTIM), Solar TerrestrialEnergyProgram (STEP) Handbooke,ditedbyR. W. Schunkp, p. 217-238,1996. Millward, G.H., R.J. Moffett, S. Quegan,and T.J. Fuller-Rowell, A Coupled Thermosphere-Ionosphere-PlasmasphMeroedel

were in the neutral winds and composition. So, if measurementsfind a stronger responsein the upper atmosphere during this or any other mid-latitude eclipseonemay,following the results presentedhere, assumethat these originate from outside the thermosphere-ionospheresystem. Studies by Fritts and Luo [1993]suggesthat perturbationsgenerated by the interrupted ozone heating during an eclipse

(CTIP), SolarTerrestriaEl nergyProgram(STEP) Handbook, edited by R. W. Schunk,pp. 239-279, 1996. Rishbeth,H., R. J. Moffett, and G. J. Bailey, Continuityof Air Motion in the Mid-Latitude Thermosphere,J. Atmos. Terr.
Phys., 31, 1035-1047, 1969.
Rishbeth,H., and I.C.F. Miiller-Wodarg, Vertical Circulationand ThermosphericComposition:A ModellingStudy,submittedto Ann. Geophys.,1999.

may propagate upwards into the thermosphere and have an important dynamical influence there. In a more extensive future study, their predicted global perturbation profile may be adapted for use at CTIP's lower boundary. In that way, eclipse disturbances originating from both below the thermosphere and from in-situ can be consideredand their

Roble, R.G., B. A. Emery, and E. C. Ridley, Ionosphericand ThermosphericResponseoverMillstone Hill to the May 30, 1994 Annual SolarEclipse,J. Geophys.Res., 91, 1661-1670,
1986.
Singh,L., T. R. Tyagi, Y. V. Somayajulu,P. N. Vijayakumar, R. S. Dabas, B. Loganadham,S. Ramkrishna, P. V. S. Rama Rao, A. Dasgupta,G. Navneeth,J. A. Klobuchar,and G. H.

relativeimportanceto the thermosphere/ionosphesryestem
evaluated.
The simulationsassumedgeomagneticallyquiet and con-

Hartmann, A Multi-Station Satellite Radio BeaconStudy of IonosphericVariations During Total Solar Eclipses,J. Atmos. Terr. Phys., 51, 271-278, 1989.

stant conditions. In reality, the high-latitude precipitation

and electric field strength vary randomly and may thus ob-
scure some of the weaker effects. No distinction was made in the CTIP simulation between
various wavelength bands in that the intensity at all wave-

I.C.F. Miiller-Wodarg,AtmosphericPhysicsLaboratory,Department of Physicsand Astronomy,University CollegeLon-
don,67-73RidingHouseStreet,LondonWlP 7PP,U.K. (e-mail: ingo@apg.phu.cl.ac.uk)

lengths was set to zero in the totality footprint. It is known that up to 10-20 % of the solar soft-X-ray and EUV radiation ((100nm) originate from the solarcorona,which is not darkenedduring an eclipse.As a result,thesewavebands will not disappear entirely.
With the limitations noted, the simulations presented

A.D. Aylward, AtmosphericPhysicsLaboratory,Department of Physicsand Astronomy,University CollegeLondon,
67-73 Riding HouseStreet, LondonWlP 7PP, U.K. (e-mail: alan@apg.phu. cl.ac.uk)
M. LockwoodS, paceScienceDepartment,RutherfordAppleton Laboratory,Chilton, Didcot, Oxfordshire,OXll 0QX, U.K. (e-mail:m.lockwood@rl.auck.)

provide us with predictions of the major changesto the

ionosphere-thermospheresystem during the August 1999 (ReceivedMay 8, 1998; revisedJuly 15, 1998; eclipse and have provided an important basis for evaluat- acceptedAugust28, 1998.)