QuantificationsofGeomagneticStormImpact

onTECandfoF2duringMarch17,2013event

NASAGoddardSpaceFlightCenterhttp://ccmc.gsfc.nasa.gov

JaSoonShim1*,Ioanna Tsagouri2,LarisaGoncharenko3M.LeilaMays1,Aleksandre Taktakishvili1,LutzRastaetter4,MariaMKuznetsova4

Modelers:Ionosphere/ThermosphereModelshostedatCommunityCoordinatedModelingCenter

1. CUA/NASAGSFC,Greenbelt,MD,USA,2. NationalObservatoryofAthens,Greece

3. MITHaystackObservatory,Westford,MA.USA4. NASA/GSFC,Greenbelt,MD,USA,

• Quantificationofstormimpact

o Quiet-timebackgroundreferences

o foF2andTECchanges

o Model/datacomparison

• ImpactsofuncertaintyintheIMFonTECsimulation

• Summary

Outline

TECandfoF2at10Ionosonde Stations• 4stationsfromUS,4fromEurope

and2fromSouthAmericatoinvestigate:o latitudeandlocaltime

dependenceo hemisphericasymmetry

• Observations:o foF2datafromtheGlobal

IonosphereRadioObservatory(GIRO)

o GPSverticalTECdatafromMITHaystackObservatory(error<4TECU)

Quiet-timeReferences

• Onedaybeforestormonset

• Fiveconsecutivedaysbeforestormonset

• Fivequietestdayswithin30dayspriortostorm

• 30dayspriortostorm

dailyaveragedKp,AE,Dst,andF10.7

• Onedaybeforestormonset

• Fiveconsecutivedaysbeforestormonset

• Fivequietestdayswithin30dayspriortostorm

• 30dayspriortostorm

Quiet-timeReferences

dailyaveragedKp,AE,Dst,andF10.7

• Onedaybeforestormonset

• Fiveconsecutivedaysbeforestormonset

• Fivequietestdayswithin30daysprior tostorm

• 30dayspriortostorm

Quiet-timeReferences

dailyaveragedKp,AE,Dst,andF10.7

• Onedaybeforestormonset

• Fiveconsecutivedaysbeforestormonset

• Fivequietestdayswithin30dayspriortostorm

• 30daysprior tostorm

Quiet-timeReferences

dailyaveragedKp,AE,Dst,andF10.7

TEC:ComparisonofFourBackgrounds

• Thebackgrounds arerepeatedacross3daysofthestormevent.

• TECononedayprior tothestorm(redline)islargerthanotherreferences.

• 30-daymedian(green)andmeanofthe5quietestdays(blue)aremoresuitable(difference<1TECU)

• Ionosphere-thermosphere modelsimulationsalsoshowsimilarfeatures(not shownhere).

StormtimeTEC30-daymedian5quietest_ave066-070_ave075(03/16)

Athens(23.5E,38.0N)

TECChangeswithrespecttothefourdifferentbackgrounds

• dTEC=100*(TEC– TEC_q)/TEC_q,whereTEC_q =quiettimebackground

• DifferenceinTECchangesbetweenredandblue (orgreen)appearsmorethan100%.

• 30-daymedian(green)andmeanof 5quietestdays(blue)aremoresuitable.

30-daymedian5quietest_ave066-070_ave075(03/16)

Athens(23.5E,38.0N)dTEC

foF2changesduringthemainphaseStation

dfoF2>20% dfoF2<-20%StartTime Max

(%)t_max Duratio

n(hrs)Starttime

Min(%)t_min Duratio

n(hrs)UT LT UT LT UT LT UT LTEuropeChilton 18.2 18.2 24% 18.2 18.2 0.3

Pruhonice 11.0 12.0 46% 11.8 12.8 3.3Ebre 11.75 12.75 97% 23.0 0.0 8.3 6.75 7.75 -31% 7.7 8.7 2.7Athens 11.5 13.5 83% 22.8 0.8 7.3

NorthAmericaIdahoNat.Lab 7.5 0.5 -45% 9.5 2.5 11

Boulder 9.5 2.5 -45% 16.0 16.0 16MillstoneHill 7.8 2.8 -48% 9.3 4.3 6.7EglinAFB 15.4 9.4 -31% 15.8 9.75 1

SouthAmericaJicamarca

Port Stanley 19.5 15.5 58% 19.5 15.5 4.5

dfoF2=(foF2–foF2_med)/foF2_med*100

• Afewhoursafterstormonset:o Europeansectorinthedaytime:positiveeffectsduetoincreases

inionizationo NorthAmericainthepost-midnightsector:negativestormeffects

causedbytheneutralcompositiondisturbance(Prölss,1993)

TECchangesduringthemainphaseStation

dTEC >50% dTEC <-40%start time

Max[%]t_max duration

(hrs)Start time

Min[%]t_min duration

(hrs)UT LT UT LT UT LT UT LTEuropeChilton 10.6 10.5 91.6% 11.5 11.5 2.8 20.2 20.1 -59.1% 11.5 11.5 1.6

Pruhonice 8.8 9.7 123.4% 11.4 12.4 5.8Ebre 9.6 9.6 144.5% 19.9 20.0 10.7Athens 8.3 9.8 148.9% 17.3 18.8 13.7

NorthAmericaIdahoNat.Lab. 9.1 1.6 209.3% 11.8 4.3 5.2

Boulder 9.5 2.5 89.4% 10.3 3.3 3.0MillstoneHill 10.5 5.7 75.1% 19.5 14.7 3.2 9.1 4.3 -43.7% 19.5 14.7 0.3EglinAFB 11.0 5.2 89.6% 19.0 13.2 2.6

SouthAmericaJicamarca 8.7 3.5 232.2% 8.9 3.8 5.1 8.3 3.1 -43.1% 8.9 3.8 0.1

Port Stanley 17.0 13.1 270.7% 20.3 16.4 2.8

• Samecolordepictssimilar latitudesanditshows similarresponses tothestorm.• BothfoF2andTECresponses tothestormarepositive phaseinEuropeansector.• NoticeabledifferencebetweenthefoF2andTECresponse inNorthAmericasector:

o TECshowsmainlypositive effects,whilefoF2showsnegativeeffects.• TECenhancementatPortStanley(42S)isaboutthreetimeslargerthanthatatEglin(40N).• AtJicamarca,foF2changes<|20%|, butTECchangegoesupto230%.

foF2andTECatBoulder

• during themainphase,o dfoF2<0o dTEC >0:contribution from

plasmasphere

foF2

TEC

dfoF2anddTEC

AssessmentofModelPredictionModel ModelSettingDescription/Modelers Lower andUpper

boundary forTECcalculation(km)EmpiricalModel

IRI 2012IRI-2012usingIRI-corrmodelfortopsideNeandusingCCIR(InternationalRadioConsultativeCommittee)forF-peakplasmafrequency foF2, DieterBilitza (GMU,NASA/GSFC)

~60 ~2,000

PhysicsBasedIonosphereModelIFM IFMdrivenbyF10.7andKp,Robert W.Schunk etal.(USU) ~90 ~1,400

SAMI3 SAMI3withtheneutralwindmodelHWM93, JosephHuba etal.(NRL) ~90 ~2,000Physics-basedCoupled Ionosphere-ThermosphereModel

CTIPE CTIPe3.2drivenbyWeimer[2005],TimothyFuller-Rowelletal.(NOAASWPC)

~140 ~2000

GITM GITM2.3drivenbyWeimer2005,AaronRidleyetal.(UM) ~90 ~600

TIE-GCM TIE-GCM2.0drivenbyWeimer[2005], R.G.Robleetal.(HAO,NCAR) ~90 ~600Physics-basedDataAssimilationModel

USU-GAIM USU-GAIM2.4.3withGPSTECobservationsfromup to400groundstations(-60° <lat <60°),Robert W.Schunk etal.(USU)

~90 ~1,400

RMSE

• AverageRMSEfor10and6stationsforTECandfoF2,respectively• ScaledTEC=TEC*(Obs_med/TEC_med)• ShiftedTEC=TEC–min(TEC_med)• DegreeofImprovementofpredictingperformancebyscalingdependsonmodels.• AveragedGPSTECerror<2TECU• 3TECU<TECRMSE<12TECU• 1.6MHz<foF2RMSE<3.6MHz

TEC foF2

where,ratio_max=dfoF2_max_model/dfoF2_max_obs(ordTEC_max_model/dTEC_max_obs)

dt_max=|t_max_obs - t_max_mod|dt_min=|t_min_obs - t_min_mod|

Model dfoF2>20% dfoF2<-20%

ratio_max dt_max ratio_min dt_min

IFM 0.99 8.25 0.71 2.95

SAMI3 0.92 4.50 1.84 2.38

CTIPE 2.54 0.00 0.78 2.67

GITM 2.42 1.00 0.60 2.12

TIE-GCM 0.97 3.5 1.24 2.92

USU-GAIM 0.84 0.88

dTEC >50% dTEC <-40%

ratio_max dt_max ratio_min dt_min

1.5 4.6 1.2 1.9

2.0 6.3 1.5 1.6

0.5 3.6 1.2 2.3

3.7 3.9

0.8 4.6 1.1

0.9 3.1 1.3 2.4

RatioofChanges

• Differencesinratio_max (anddt_max)amongmodelsarelargerthanthoseinratio_min (anddt_min).

• Red:betterratio• Blue:bettertimeprediction

ImpactsofUncertainty

IntheInterplanetaryMagneticField(IMF)

onTEC

SolarWindParametersfromEnsembleofWSA+ENLIL+ConeModelruns

|Vx|

IMFClockangleensembleofITmodel runs:90°,135°,and180°

ACEdatapredictedmeanvalueoutofensemble(generatedbeforetheevent)bestfitoutofensemble(aftertheevent)

Shockarrivaltime:• predicted:2013-03-16T21:34Z• adjusted: 2013-03-17T05:12Z• about7hourdifference

IMFClockAngleEnsemble

• UncertaintyinIMFclockangle(w/adjustedsolarwindparameters)hasnoticeableimpactonTECinmidlatituderegionduringthemainphase.

GPSvTECw/ACEw/ENLIL(90°)w/ENLIL(135°)w/ENLIL(180°)

CTIPe TIE-GCM

Summary• QuantifiedstormimpactsonfoF2andTECat10selectedionosonde locations.• Comparedfourdifferentquiet-timereferences:

o 30-daymedianandmeanoffivequietestdaysarecomparable.o averaged5consecutivedaysandonedaybeforethestormmaynotbesuitable.

• Duringmainphase,o Europeansector:both foF2andTECresponsetothestormarepositivephaseo NorthAmericasector:foF2shows negativeeffects,whileTECshowspositiveresponse.Itis

possiblyduetoplasmaspherecontribution.o TECenhancementatPortStanley(42S)isaboutthreetimeslargerthanthatatEglin(40N).

• EvaluatedhowwellIonosphere-thermospheremodelsreproducetheTECandfoF2changesduringthemainphase.o RMSEofthemodelsislargerthanerrorsinobservations.o performancedependsonmetricsandquantitiesselected.

• UncertaintyinIMFclockanglehasnoticeableimpactonTECinmidlatituderegionduringthemainphase.