Flowout_ver0.1.pdf

JOURNAL OF GEOPHYSICAL RESEARCH, VOL. ???, XXXX, DOI:10.1029/,

Effects of changing solar wind conditions on open drift paths.Center for Space Engineering, Utah State University

Abstract. We investigate the rate at which the open drift paths in the near earth mag-netosphere convert to closed paths and thus find the rate of flow out losses from the in-ner magnetosphere. The Block-Adaptive-Tree Solar-Wind Roe-Type Upwind Scheme, (BATS-R-US) model along with the Rice Convection Model (RCM) and the Fok Ring Current(FRC) model available at Community Coordinated Modeling Center (CCMC) are usedto evaluate the effects of changing solar wind conditions on particle drifts. Geomagneticstorms which have been classified as category-I storms are analyzed by the numericalmodels and their performance is compared against simultaneous ground magnetic mea-surements using movie maps. The results indicate that ground magnetic disturbance re-mains asymmetric for some time after the start of recovery phase even for a category-I storm. FRC simulation results suggest that that the flow out losses reduce under weak-ened magnetospheric convection. The contributions from the magnetopause currents andmagnetotail currents can be used explain some of the discrepancies observed betweenthe movie maps and FRC simulation values.

1. Introduction

Growth of partial ring current and the formation of theclosed ring current are one of the most important features ofa geomagnetic storm. The intensity of a storm is measuredby the Dst (disturbance storm time) index, which is a proxyfor the strength of the ring current. It has been observedthat the Dst index generally has a two phase decay as themagnetosphere recovers from a storm. It has been shownthrough simulations that the fast initial decay of the Dstindex is caused by the flow out loss of energetic ring cur-rent ions through the magnetopause [Takahashi et al., 1990;Liemohn et al., 2001a]. This is followed by slower chargeexchange mechanisms in the later recovery phase. The ini-tial fast flow out loss is controlled by the rate of cessationof solar wind driving.

The initial decay rate of storms with abrupt cessation ofsolar wind driving and other storms with gradual recoverywas compared by OBrien et al. [2002]. They used 29 stormsin the period Nov. 1963 - September 2001. Their findingssuggested that the storms with abrupt northward turningof the IMF Bz show the same recovery in the first 6 hoursor slightly more recovery than do the storms with gradualnorthward turnings. In another work, Patra et al. [2011]studied 13 storms in the period 2001 -2007, with abruptnorthward turning of the IMF Bz after the peak in Dstindex was observed. They concluded that the two phase de-cay of the Dst index was still evident in these storms eventhough the solar wind driving was turned off.

Both these works agree on the fact that the rate of re-covery is not affected significantly by the northward turning(i.e., shutoff of convection). OBrien et al. [2002] suggestthat the flow-out provides an additional loss mechanism,being equal to or greater than charge-exchange loss duringslow-shutoff-storm recovery causing the Dst index to havesimilar recovery times. In Patra et al. [2011] ’s work, it wasreported that modeling the Dst index by including contribu-tions from other magnetospheric currents, most noticeablythe tail current provided a high degree of fidelity in estimat-ing Dst. The plasma sheet density too has an importantrole to play in the ring current build up and decay as shownby Liemohn and Kozyra [2005].

Copyright 2013 by the American Geophysical Union.0148-0227/13/$9.00

Kozyra et al. [2002] have successfully modeled stormswith northward turning of the IMF Bz field using empiri-cal and numerical models [Kozyra and Liemohn, 2003]. Theenergy transfer from the solar wind and the resulting con-vection electric field plays an important role in the fast flowout losses of ring current ions in these models.This convec-tion electric field is closely related to the interplanetary E-field Gonzalez et al. [1989]. The various solar wind magne-tospheric coupling functions represent this relationship. Aclear consensus does not yet exist as to which function de-scribes the convections electric field the best. This leads todifficulty in interpreting model results related to the ener-gization of the ring current and its subsequent loss processes.

In this work we compare ring current simulation resultsagainst global low latitude magnetometer data for a stormwith abrupt northward turning of the IMF Bz (hence, lead-ing to cessation of solar wind driving). The selected stormhas been classified as a category I storms in Patra et al.[2011]. Category-I storms have been classified on the ba-sis of their similar performance under different solar windcoupling functions. Magnetometer data from various lowlatitude stations are plotted in a unique polar coordinateconfiguration first proposed by Love and Gannon [2010] tocompare against ring current simulations.

In the next section we explain the models available atCCMC, which have been used for this analysis. Next in sec-tion 3 we explain the procedure used to generate the moviemaps from the magnetometer data. A comparison of theresults of the simulation with the magnetic data is done sec-tion 4.

2. Models at CCMC

The Community Coordinated Modeling Center (CCMC)Runs-on-Request System (RoR System) is used to obtainoutput from the BATS-R-US global MHD model run alongwith the Rice Convection model (RCM) Toffoletto et al.[2003] and the Fok Ring Current model of the inner magne-tosphere model Buzulukova et al. [2010]; Fok et al. [2001].

The BATS-R-US code solves the governing equations ofmagnetohydrodynamics Zeeuw et al. [2004]. All terms de-scribing deviations from ideal MHD are included throughappropriate source terms. The governing equations for anideal, non relativistic, compressible plasma may be writtenin a number of different forms. In primitive variables, thegoverning equations, which represent a combination of the

1

X - 2 SPENCER ET AL.: RATE OF FLOW OUT LOSSES.

Euler equations of gas dynamics and the Maxwell equationsof electromagnetics, may be written as:

∂ρ

∂t+ u.∇ρ+ ρ∇.u = 0 (1)

ρ∂u

∂t+ ρu.∇u+∇p− j×B = 0 (2)

∂B

∂t+∇×E = 0 (3)

∂ρ

∂t+ u.∇p+ γp∇.u = 0, (4)

where the current density j and the electric field vectorE are related to the magnetic field B by Amperes law andOhms law, respectively:

j =1

µ0

∇×B (5)

E = −u×B (6)

The Comprehensive Ring Current Model(CRCM) couplesthe Rice Convection Model(RCM) and the kinetic model ofFok and coworkers Fok et al. [2001]. The calculations areperformed in two steps. First, the evolution of distributionfunction at each point is calculated which is due to driftand losses (FokRC model). Then, the field-aligned currentsin the ionosphere and ionospheric potential are calculatedusing RCM scheme (for the details, see Fok et al. [2001]).Field-aligned currents are calculated from a current conti-nuity equation between the magnetosphere and ionosphereBuzulukova et al. [2010]:

J||i =1

ricos2λ

∑

i

(

∂ηj∂λ

∂Wj

∂φ− ∂ηj

∂φ

∂Wj

∂λ

)

(7)

where the summation is done at fixed λ, φ point and overallM,K points, J||i is a sum of ionospheric field- aligned cur-rent densities for both hemispheres, Wj is the kinetic energyof a particle with given λ, φ,M,K and ηj is the number ofparticles per unit magnetic flux (density invariant in termsof RCM) associated with ∆M,∆K:

ηj = 4√2πm

3/20

fs(λ, φ,M,K, )M1/2∆M∆K. (8)

Using the distribution of field-aligned currents, the iono-spheric potential is obtained from equation (8). We assumehere that Bi is the same for both hemispheres. By definition,J||i here describes only Region II field aligned currents.

3. Moviemaps

The Dst index is used as an indicator of the ring cur-rent energy. It is calculated from a weighted average ofdisturbance data from a sparse longitudinal distribution of4 low latitude magnetic observatories. However Dst doesnot measure the local time distribution of the low-latitudemagnetic disturbances. The availability of an extensive net-work of high quality magnetometer stations provides us theopportunity to study the local time distribution of magneticdisturbances. Love and Gannon [2010] were the first to useunique polar coordinate maps of the local time functionaldependence of storm time disturbance. They plotted themagnetic disturbance data in a geometry that resembled thephysical structure of the ring current. These “movie maps”permit detailed inspection of the data, their variation intime and their variance in space.

Another popular scheme to analyze magnetometer data isto create panoramic views by making contour plots of mag-netic disturbance across a domain of local time and univer-sal time (LT-UT) Zaitev and Bostrom [1971]; Clauer and

McPherron [1974]; Clauer et al. [2003]. We use maps sim-ilar to the movie maps in addition to the LT-UT plots tocompare the FRC results with the ground magnetic distur-bances.

We use a technique similar to the that used for the gen-eration of Kyoto Dst index to calculate the magnetic dis-turbances. We find the two quietest days in a month andtake the average to create the quiet time B-field (StatQavg).This is subtracted from the magnetic field measured at eachmagnetometer station (StatH) as shown in eqn. 9.

DistH =StatH − StatQavg

cos(φ); (9)

The latitudinal correction is accounted by dividing thecosine of the each station’s magnetic latitude value(φ).DistH(θm) is a smooth curve generated by interpolatingeach magnetic station data using a Fourier series fit to thedata as shown in eqn. 10. The various model parameters(DstO, a

ci , a

si ) are obtained with a least squares algorithm.

The movie maps are plotted in a polar co-ordinate systemsimilar to the ones used by Love and Gannon [2010].

DistH(θm) = DstO +3

∑

i=1

acicos

(

2πiθm1440

)

+

3∑

i=1

asi sin

(

2πiθm1440

)

, (10)

where DstO is a representive Dst generated from all thelow-latitude stations under consideration.

The decomposition in terms of Fourier terms is motivatedby a need for a complete basis set that is periodic in localtime. The reason for only choosing Fourier expansion up todegree 3 is guided by the need to satisfy the spatial Nyquistcriteria [Clauer and McPherron, 1974]. It was found byClauer and McPherron [1974] that the distribution of mag-netic observatories around the world at that time was in-adequate to define coefficients beyond the third harmonic,which corresponds to sine waves of 8-hour period in localtime and by the Nyquist criteria requires a separation of atmost 4 hours. Love and Gannon [2010] used data from 20and 25 observatories for each of the storms that they an-alyzed. The magnetic data for this work was downloadedfrom INTEMAGNET consortium which has validated andpublished 1 min resolution digital data from the memberobservatories. We use data from 51 low and mid-latitudestations (max. latitude = 61.8◦, min. absolute latitude= 4.3◦ ) for higher time resolution.

The polar plots used in the generation of the movie mapsaid in better visual understanding of the ring current systemunder observation. Each instantaneous disturbance valueDistH from each observatory is plotted radially, where thezero value is on a black circle centered at the origin. Thispermits unambiguous plotting of disturbance data that arepositive (inside the zero-value circle) and negative (outsidethe circle). The azimuthal angle used for plotting eachDistH value is the local magnetic time for the observatory.

Selected movie maps will be made freely available on theweb. In the next section we will compare the FRC resultsand the movie maps for the chosen category I storm.

4. Simulation Results and Moviemaps

In Spencer et al. [2011] storms with abrupt turning ofIMF Bz in the early recovery phase, have been classified

SPENCER ET AL.: RATE OF FLOW OUT LOSSES. X - 3

−100

−50

0

50

Sym

H(n

T)

20406080100120

Asy

mH

(nT

)

0

500

1000

−A

L(nT

)

−20

0

20

Bz(

nT)

−600

−500

−400

Vx(

Km

/s)

228 228.2 228.4 228.6 228.8 229 229.2 229.4 229.6 229.8 2300

20

40

60

80

−N

p(#/

cc3 )

Figure 1. The magnetic indices and the solar wind con-ditions during the Category I storm starting on day 228,year 2001. The dashed vertical lines correspond to theeight instances of time discussed in the text.

into category I and II, depending on the similarity or dis-similarity of their coupling functions and resultant analysisby the WINDMI model. Category-I storms have similarsolar wind coupling function values leading to similar Dstestimation by the WINDMI model irrespective of the inputcoupling function used. These storms are ideal in analyzingthe ring current response in the recovery phase of a geomag-netic storm. Since the various coupling functions estimatesimilar low energy input during the recovery phase of thestorm, we expect the results of this simulation to hold irre-spective of which solar wind driving formula is used.

Fig. 1 shows the solar wind conditions and theSymH,ASymH, and AL magnetic indices for the categoryI storm starting on day 228 (Sept.), year 2001. A suddenstorm commencement started around 11:00 Hrs on Sep, 17.This was followed by the southward turning of the IMF Bzand the main phase of the storm. At 21:20 Hrs on Sep. 17,the IMF Bz suddenly turns northward triggering the recov-ery of the storm. Note that the IMF Bz turns northwardright after the peak in the SymH index was observed. Itwas shown earlier in Spencer et al. [2011], that four othercoupling functions estimated similar low energy input as cal-culated by the rectified V Bs (B-southward) function, in therecovery phase of this storm.

We choose 8 instances in time to illustrate a few uniqueobservations made for this storm. The first instance is at

Figure 2. Bottom. UT-LT map showing the azimuthal(LT) variation of the midlatitude geomagnetic distur-bance on day 228, 2001. Top. The SymH index is for thestorm duration plotted. The dashed vertical lines corre-spond to the eight instances in time chosen for this study.

10:00 hrs on Sep. 17 representing the quiet time featuresbefore the start of the storm. Three other instances are cho-sen in the initial phase of the storm showing some symmetricand asymmetric features during this phase. Two instanceseach in the main phase and recovery phase are chosen toshow the how the asymmetric response in the main phasechanges to symmetric during the later part of the recoveryphase. These eight instances are shown by the dashed ver-tical red lines in figure 1.

A popular way of representing the temporal and spa-tial (azimuthal) variations of the horizontal component ofthe magnetic disturbance H is to display them in a two-dimensional LT-UT diagram. The usefulness of the LT-UTmap consists of the identification of its features (the spatiallocation of the field disturbance) with specific current sys-tems [Daglis et al., 2003]. The LT-UT plot for this storm isplotted in figure 2 (bottom). It shows the azimuthal (LT)variation of the low/mid-latitude geomagnetic disturbancesobserved at the magnetometer stations. For reference wealso plot the SymH index at the top of this figure.

Figure 3. Comparison of Fok Ring current Simulationresults for the category I storm starting on day 228, year2001 with magnetospheric indices and movie maps. Thefirst row plots the SymH and the FRC proton energy.The second and third rows show the AL and AsymHvalues for the storm. Comparison of the FRC ion fluxand the movie maps is shown in the fourth row at 10:00,and 11:20 UT.



The observed diagonal trend in the data having 45 de-grees of gradient is the effect of the Earths rotation whichmatches the station motion across the map. Clauer et al.[2006] investigated this observation and suspected that it re-sults because peaks in the disturbance profile are observedby stations in a specific region, but they rotate with timecarrying the peak in the profile with them. A solution toeliminate this effect was also proposed by them where a ref-erence time was selected and subtracted from the data valuesat the reference time from all later values for each station.We have however not used this method in producing fig. 2.

Before the start of the storm just before 228.5, the dis-turbance measured at the magnetometer stations are sym-metric and show almost negligible disturbance, as referencedby the first vertical line in fig 2. The start of the storm issignaled by the arrival of a solar wind pressure pulse trigger-ing the sudden storm commencement (SSC). This is shownby the second vertical dashed line in fig 2 . It can be seenthat the positive disturbance due to the compression of themagnetosphere is observed at all local times. This is due tothe fact that the IMF Bz is still northward. The next twovertical lines signify the start of the magnetospheric convec-tion while the effect of magnetospheric compression is stillactive.

The signature of convection is localized first in the nightside and gradually its magnetic signature spreads to coverthe entire earth (vertical lines 5 and 6). The next instance intime chosen during the early recovery phase (vertical line 7)shows the response of the magnetometer station to a pres-sure pulse during the early recovery phase. This can also beseen as a sudden recovery in the SymH index. The strong,dynamic azimuthal variations of the midlatitude disturbanceprovide qualitatively different information from the placid,simpler view of the storm afforded by SymH or even theAsym index. The LT-UT plots provide a different perspec-tive as compared to the SymH index and when analyzedwith the movie maps, a lot of new information can be ob-tained.

We ran the BATSR-US model along with RCM and FRCat CCMC for the category - I geomagnetic storm starting onday 228, year 2001. The ring current particle flux from thesimulation are overlaid on the corresponding movie maps asshown in figs. 3-6. Each of the figures from figs. 3-6 havefour rows each. The first row in each figure plots the SymHfor the entire storm. Also plotted in the first row is the totalproton energy calculated by the FRC. The ring current en-ergy Wrc can be related to the Dst/SymH index using theDessler-Parker-Schopke (DPS) relation [Dessler and Parker ,1959; Sckopke, 1966]:

Dst =µ0Wrc(t)

2πBER3

E

(11)

where Wrc is the plasma energy stored in the ring currentand BE is the earth’s surface magnetic field along the equa-tor.

The second and third rows plot the ASymH and AL in-dices for the storm respectively. The fourth row comparesthe FRC simulation results with the magnetic data from themovie maps. In each of the figures, two instances in timeare compared in the fourth row. Each image from the FRCplots the calculated flux of protons in the inner magneto-sphere. The plots are color coded in a 2-D surface plot. Themovie map plots the magnetospheric disturbances in a polarcoordinate. A central black circle designates the zero dis-turbance reference. Magnetospheric data from each stationis plotted as red dots and a smooth fit according 10 is plot-ted to represent the local time variation of the disturbance.The spatial scales of the disturbance calculated by FRC areaccurate while the movie map data is just illustrative.

We have used the particle flux with energy 37.7 KeV tocompare against the movie maps. Liemohn et al. [2001b]

found that the average energy is 40 KeV during the mainphase of three storms that they analyzed. It is thus expectedthat the dominant contribution to the ring current in themain phase and early recovery phase comes from particelswith energies around 40 KeV. We discuss the developmentof particles with other significant energies at the end of thissection.

The left panel in the fourth row of fig. 3 shows the re-sponse of the FRC as well as the moviemap before the stormhas started (quiet time response). It can be clearly seen thatthe FRC flux of protons is symmetric. The magnetic dis-turbances too are symmetric and hardly show any deviationfrom the zero disturbance circle.

4.1. Initial Phase

Figure 3 plots the response of the FRC and movie mapbefore the storm has started and just after the sudden stormcommencement(SSC). The right panel in the fourth row ofthe figure shows the results at time 11:20 UT on day 228,year 2001. The SymH data in the first row shows a positivedisturbance. Corresponding disturbances are also seen inthe AL and ASymH indices. The moviemap result showsalmost symmetric positive disturbance across all the mag-netic stations. This is consistent with the theory of Chap-man and Ferraro [1930], the movie-map for this storm makesit clear that the onset of the initial phase is caused by anenhancement of solar wind pressure. This pushes the mag-netopause in toward the Earth and intensifies the eastwardelectric currents of the magnetopause. By Ampres law, themagnetopause currents generate a northward magnetic dis-turbance, and since the dimension of the magnetopause ismuch larger than the diameter of the Earth, positive mag-netic disturbance is seen more or less uniformly at all localtimes; the curve fitted to the disturbance data is relativelysymmetrical Love and Gannon [2010]. The FRC responseshows an increase in the particle flux on the dayside in re-sponse to this pressure pulse.

The next figure 4 illustrates the additional informationthat is obtained from the use of movie maps. The asym-metries observed during the initial phase of this storm areclearly visible in the fourth panel of this figure. The panelon the left in the fourth row plots data-simulation results attime 12:40. Compared to the right panel in the fourth rowof fig. 3 the disturbance of this initial phase becomes more



asymmetrical. This is possibly in response to mild magneto-spheric convection commencing with intermittent Bz southand connection of the interplanetary magnetic field onto thegeomagnetic field. The IMF and solar wind parameters canbe found in figure 1. Thus the obvious energization of thering current prior to the start of the main phase of the stormcan be observed through the movie maps.

The second instance that is plotted in figure. 4 is at14:20 UT. At this time the maximum asymmetry in the lat-itudinal magnetic disturbance during the initial phase of thestorm was observed, as shown in the ASymH index. Theasymmetry is created by both positive and negative distur-bances. It can be hypothesized that these represent a super-position of disturbance sustained by magnetopause currents,supported by solar wind pressure, and partial ring currents(and, even, field-aligned currents) Love and Gannon [2010];Siscoe [2006].

4.2. Main Phase

The dawn dusk asymmetry observed in magnetic stationdata has been historically interpreted as due to the strongpresence of the partial ring current. The cause of the asym-

Figure 6. Comparison of Fok Ring current Simulationresults for the category I storm starting on day 228, year2001 with magnetospheric indices and movie maps. Thefirst row plots the SymH and the FRC proton energy.The second and third rows show the AL and AsymHvalues for the storm. Comparison of the FRC ion fluxand the movie maps is shown in the fourth row at 22:00UT on day 228 and 04:20 UT on day 229.

−100

−50

0

50

Sym

H(n

T)

−20

0

20

Bz(

nT)

−8

−6

−4

−2

Cum

ulat

ive

FR

C e

nerg

y lo

ss (

1e31

J)

FlowoutCharge exchangeTotal

228.5 228.6 228.7 228.8 228.9 229 229.1 229.20

2

4

6

8

10

FR

C e

nerg

y lo

ss(1

e29

J)

FlowoutCharge exchangeTotal

Figure 7. The bottom row shows the absolute energygain of the ring current particles as simulated by the FRCmodel for the storm. The third row shows the total en-ergy gain. SymH and IMF Bz are shown in the top tworows. The dashed vertical lines correspond to the eightinstances of time discussed in the text.

metry is a combination of forces due to magnetic field gra-dients and convective electric fields. This results in a con-centration (reduction) of ion drift lines of trajectory in thedusk (dawn) magnetosphere [e.g., Takahashi et al. [1990];Liemohn et al. [2001a]], or, equivalently, a dusk-centeredpartial ring current. Figure 5 shows results at time 19:20and 21:20 UT. These times correspond to the peak asymme-try and the peak intensity observed during the main phaseof the storm. It can be seen from the movie map as wellas FRC data that the flux and the associated disturbance islarge during this time. The solar wind and IMF conditionsindicate that conditions favorable for strong convection werepresent.

At the end of the main phase approximately around 21:20UT the IMF Bz turns northward almost abruptly and therecovery of the storm is triggered. It is expected that theflow out losses which were dominant during the main phasewill be become less important in the recovery phase.

4.3. Recovery Phase

This storm starting on day 228, year 2001 was initiallychosen for study by Patra et al. [2011] since a sudden north-ward turning of the IMF Bz was observed after the peak inDst index was reached. It was hypothesized that this sud-den northward turning of the IMF Bz will lead to trappedparticle in the earth’s ring current and the flow out losseswhich were one of the dominant modes of ring current par-ticle loss during the main phase will be less important. TheFRC simulation results seem to agree with this assumptionas can been seen in fig. 6. The FRC simulation results at theinstances of time show that the peak of intensity has clearlyshifted from the nightside to the dayside(compare left paneland right panel in the fourth row). This can be possibly beexplained as a result of drift of trapped particles on closeddrift paths in the absence of convection from the nightsidein response to the northward turning of IMF Bz.

At 22:00 UT a sudden drop in the value of SymH wasobserved. A simultaneous sudden increase in solar wind dy-namic pressure was also observed as can be seen fig. 1. Thecorresponding movie map plot shows an interesting almosttriangular disturbance. This could possibly be a responseto the sudden compression of the dayside magnetosphere inresponse to the pressure pulse. The effects of the alreadyincreased asymmetric ring and possibly tail current weresuperposed with this increase in the magnetopause current.

In the late recovery phase, at 04:20 UT, day 229, year2001 an enhanced symmetric ring current is observed in theFRC simulation results. The movie map too shows a sym-metric but negative disturbance across the magnetometerdata. In this phase the particles are already trapped on theclosed drift paths while losing energy due to charge exchangewith neutral atoms.

During this storm the IMF Bz turns northward at around228.9 right after the peak in SymH was obtained. In Patraet al. [2011] it was found that various different solar windmagnetosphere coupling functions produced similar low en-ergy transfer values in the recovery phase when the IMF Bzis northward. We can obtain the total energy gain and lossfor the ring current particles from the FRC model. In figure7 we have plotted the total energy gain as well as the abso-lute energy gain of the ring current particles. The top tworows show the SymH and Bz values. The eight instancesof time chosen earlier are shown by vertical dashed lines. Itcan be clearly seen that the charge exchange losses are muchsmaller as compared to the losses due to flow out of parti-cles from the magnetosphere in the main and early recoveryphase. The ratio of the flow out to charge exchange exceeds1 only in the late recovery phase. Although the flow outlosses do not abruptly stop after the northward turning of


Figure 8. The ring current total particle flux data for energies 22.4 ,37.7, 63.3 and 106 keV from theFRC simulation are plotted.

the IMF Bz, but the flow out losses reduce sharply after thenorthward turning. For this storm the flow out losses reducedrastically within 2 hours, so for this storm it can be safelyassumed that the early recovery has a smaller contributionfrom the flow out loss.

Mitchell et al. [2001] have used ENA images of the Earth’sinner magnetosphere to compare the ring current morphol-ogy during the Bastille day event and a moderate event onJune 10, 2000 for which the IMF Bz gradually turned north-ward. The IMF Bz turns northward soon after the peakin SymH for the Bastille day event. They confirmed thatthe contribution to the ring current in the small, June 10storm and associated substorms was much further away fromEarth, and much more dependent on open drift path dynam-ics, than in the larger Bastille storm where the ions con-tributing to Dst drifted primarily on closed paths. Particlestrapped in the ring current once the magnetospheric con-vection weakens drift around earth and lose energy throughthe charge exchange process. This was seen in the ENAimages for the Bastille day storm ?. In figure 8 we plotthe ring current particle flux data from the FRC simulationfor 22.4 ,37.7, 63.3 and 106 keV energies. It can be seenthat the in the main phase the ring current was enhancedin the night side under continuous magnetospheric convec-tion. The location of the enhancement did not change muchwhen the IMF Bz was southward. For each of the energylevels, when the IMF Bz turns northward and the particlesare trapped, it can be seen that the particles drift aroundearth and eventually lose energy.

5. Discussion and Conclusion

In the previous section a particular category I storm wasanalyzed using the FRC model. The results were comparedwith a movie map created from low latitude stations spreadacross LT. The total energy of the Earth’s ring current cal-culated by the FRC and they are compared with the SymHindex. These can be seen in the first row of any of the fig-ures from fig. 3-6. Good agreement was found between thetwo in the initial and early part of the main phase. FRCmodel predicted a delayed peak in the ring current energyas compared to the values suggested by SymH index. Therecovery phase of the storm also showed a difference in thedecay time estimated by the FRC and the SymH index.

Liemohn and Kozyra [2005] tested the hypothesis thatthe observed two phase decay of the Dst, SymH indices canbe caused charge exchange processes alone. It was shownthat a two-phase decay (a sharp transition between fast andslow recovery rates) of the ring current total energy contentis produced when the plasma sheet density is dramaticallyreduced several hours prior to a sudden reduction in themagnetospheric convection strength. The reverse situation,a convection strength reduction prior to a plasma sheet den-sity decrease, does not produce a two-phase decay signature.A two-phase decay is not visible in the results for simulta-neous reduction of these two input parameters.

In our previous work [Patra et al., 2011] we have showntwo phase decay was observed for category I storms as well.


It was also observed by Liemohn and Kozyra [2005] thatthe flow out losses directly follow the convection E-fieldstrength. We observed similar results during the courseof this work. These results seem to suggest that duringthe early phase of the category I storms the contributionfrom other magnetospheric currents like the cross tail cur-rent might be important. The comparisons made for thisstudy need to be extended to other storms to validate thesesuggestions

The work done during the course of this study is an ex-ercise in data model validation using the vast network ofmagnetometer data. Interesting observations were made us-ing the unique movie maps generated from individual lowlatitude magnetometer station. Although some well knownphases of a geomagnetic storms were reliably reproduced inboth FRC results and the moviemaps, some interesting dif-ferences too were highlighted. The use of several magneticdata to validate model results gives the scientific commu-nity with a reliable multipoint tool to match against theirmodels.

Acknowledgments. Simulation results have been providedby the Community Coordinated Modeling Center at GoddardSpace Flight Center through their public Runs on Request system(http://ccmc.gsfc.nasa.gov). The CCMC is a multi-agency part-nership between NASA, AFMC, AFOSR, AFRL, AFWA, NOAA,NSF and ONR. The [MODEL NAME HERE] Model was devel-oped by the [MODEL DEVELOPER(S)/GROUP] at the [INSTI-TUTION].

The results presented in this paper rely on data collected atmagnetic observatories. We thank the national institutes thatsupport them and INTERMAGNET for promoting high stan-dards of magnetic observatory practice (www.intermagnet.org).

References

Buzulukova, N., M. Fok, A. Pulkkinen, M. Kuznetsova, T. E.Moore, A. Glocer, P. C. Brandt, G. Tth, and L. Rasttter(2010), Dynamics of ring current and electric fields in theinner magnetosphere during disturbed periods: Crcmbats-r-us coupled model, J. Geophys. Res., 115 (A05210), doi:doi:10.1029/2009JA014621.

Chapman, S., and V. C. A. Ferraro (1930), A new theory of mag-netic storms, Nature, 126, 129–130.

Clauer, C. R., and R. L. McPherron (1974), Mapping local timeuniversal time development of magnetospheric substorms us-ing mid-latitude magnetic observations, J. Geophys. Res., 79,2811–2820.

Clauer, C. R., M. W. Liemohn, J. U. Kozyra, and M. L. Reno(2003), The relationship of storms and substorms determinedfrom mid-latitude ground-based magnetic maps, in distur-bances in geospace: The storm-substorm relationship, Geo-phys. Monogr. Ser., 142, 143–157, aGU, Washington, D. C.

Clauer, C. R., X. Cai, D. Welling, A. DeJong, and M. G. Hender-son (2006), Characterizing the 18 april 2002 storm-time saw-tooth events using ground magnetic data, J. Geophys. Res.,111 (A04S90), 1208, doi:doi:10.1029/2005JA011099.

Daglis, I. A., J. U. Kozyra, Y. Kamide, D. Vassiliadis, A. S.Sharma, M. W. Liemohn, W. D. Gonzalez, B. T. Tsuru-tani, and G. Lu (2003), Intense space storms: Critical is-sues and open disputes, J. Geophys. Res., 108 (A5), 1208, doi:doi:10.1029/2002JA009722.

Dessler, A., and E. N. Parker (1959), Hydromagnetic theory ofgeomagnetic storms, J. Geophys. Res., 64, 2239–59.

Fok, M. C., R. A. Wolf, R. W. Spiro, and T. E. Moore (2001),Comprehensive computational model of earth’s ring current,J. Geophys. Res., 106 (A5), 8417–8424.

Gonzalez, W. D., B. T. Tsurutani, A. L. Gonzalez, E. J. Smith,F. Tang, , and S.-I. Akasofu (1989), Solar wind-magnetospherecoupling during intense magnetic storms (1978-179), JGR,94 (A7), 8835–8851.

Kozyra, J. U., and M. W. Liemohn (2003), Ring current energyinput and decay, Space Science Reviews, 109 (A8), 105–131.

Kozyra, J. U., M. W. Liemohn, C. R. Clauer, A. J. Ridley, M. F.Thomsen, J. E. Borovsky, J. L. Roeder, V. K. Jordanova, andW. D. Gonzalez (2002), Multistep dst development and ringcurrent composition changes during the 46 june 1991 magneticstorm, JGR, 107 (A8), 1295–1322.

Liemohn, M. W., and J. U. Kozyra (2005), Testing the hypothe-sis that charge exchange can cause a two-phase decay, in TheInner Magnetosphere: Physics and Modeling, Washington DCAmerican Geophysical Union Geophysical Monograph Series,vol. 155, pp. 211–+.

Liemohn, M. W., J. U. Kozyra, C. R. Clauer, and A. J. Ridley(2001a), Computational analysis of the near-earth magneto-spheric current system during two-phase decay storms, JGR,106 (A12), 29,531–29,542.

Liemohn, M. W., J. U. Kozyra, M. F. Thomsen, J. L. Roeder,G. Lu, J. Borovsky, and T. E. Cayton (2001b), Dominant roleof the asymmetric ring current in producing the stormtimedst*, JGR, 106, 10,883–10,904.

Love, J. J., and J. L. Gannon (2010), Movie-maps of low-latitudemagnetic storm disturbance, Space Weather, 8 (S06001), doi:doi:10.1029/2009SW000518.

Mitchell, D. G., K. C. Hsieh, C. C. Curtis, D. Hamilton, H. D.VOSS, E. C. Roelof, and P. C:son-Brandt (2001), Imaging twogeomagnetic storms in energetic neutral atoms, Geophys. Res.Lett., 28 (6), 1151–1154, doi:doi:10.1029/2003JA010164.

OBrien, T. P., R. L. McPherron, and M. W. Liemohn (2002),Continued convection and the initial recovery of dst, Geophys.Res. Lett., 29 (23), 2143, doi:doi:10.1029/2002GL015556, 2002.

Patra, S., E. Spencer, W. Horton, and J. Sojka (2011), Studyof dst/ring current recovery times using the windmi model, J.Geophys. Res., 116 (A02212), doi:10.1029/2010JA015824.

Sckopke, N. (1966), A general relation between the energy oftrapped particles and the disturbance field near the Earth,J. Geophys. Res., 71, 3125.

Siscoe, G. L. (2006), Global force balance of region 1 current sys-tem, Journal of Atmospheric and Solar-Terrestrial Physics,68, 2119–2126.

Spencer, E., P. Kasturi, S. Patra, W. Horton, and M. L. Mays(2011), The influence of solar wind-magnetosphere couplingfunctions on the dst index, J. Geophys. Res., 116 (A12235),doi:doi:10.1029/2011JA016780.

Takahashi, S., T. Iyemori, and M. Takeda (1990), A simulation ofthe storm time ring current, Planet. Space Sci., 38 (9), 1133–1141.

Toffoletto, F., S. Sazykin, R. Spiro, and R. Wolf (2003), In-ner magnetospheric modeling with the rice convection model,Space Science Reviews, 107, 175–196.

Zaitev, A. N., and Bostrom (1971), On methods of graphical dis-playing of polar magnetic disturbances, Planet. Space Sci., 19,643–649.

Zeeuw, D. L. D., S. Sazykin, R. A. Wolf, T. I. Gombosi, A. J. Rid-ley, and G. Toth (2004), Coupling of a global mhd code and aninner magnetospheric model: Initial results, J. Geophys. Res.,109 (A12219), 8417–8424, doi:doi:10.1029/2003JA010366.

Documents

Flowout_ver0.1.pdf