1 Introduction

The people’s understandings for the mech-anism and framework of space weather fore-casting are often based on the analogies withterrestrial weather forecasting. However, thereare a number of significant gaps between bothtypes of weather forecasting. The major differ-ence is whether the system for weather fore-casting can be approximated by a closed sys-tem. In the case of terrestrial weather forecast-ing, the system can be practically approximat-ed by a closed system to predict its future sta-tus from initial values. This is the reason whythe global atmospheric circulation model can

be applied for long-term weather forecasting.In the case of space weather forecasting, thecauses of disturbances in space environmentaround the earth (called geospace) originatedfrom the solar activity and the upstream solarwind. If there is no information from the solaractivity and the upstream solar wind, thefuture condition of geospace would be unpre-dictable, even with an excellent numericalweather forecasting model is provided foroperation［1］. Therefore, numerical simula-tion［2］and data assimilation technology thatprovides an integrated solution to handlinginformation about the solar activity and solarwind variations and its propagation from the

41NAGATSUMA Tsutomu et al.

2-1-4 Preceding Monitoring of Solar Wind Towardthe Earth Using STEREO

NAGATSUMA Tsutomu, AKIOKA Maki, MIYAKE Wataru, and OHTAKA Kazuhiro

Acquisition of solar wind information before it reaches the earth is a key to extending thelead time for predicting disturbances in space environment around the earth (called geospace).Since the geospace is an open system, the future status of geospace is unpredictable withouthaving an uninterrupted supply of information about the upstream of the solar wind, as a drivingforce for the geospace disturbanes. STEREO (Solar TErrestrial RElations Observatory) is theNASA’s scientific mission involving two spacecrafts traveling slowly away from the earth basical-ly on the same orbit as the earth to observe the Sun and the solar wind in a stereoscopic view.Like the ACE (Advanced Composition Explorer) mission, STEREO broadcasts real-time beaconsignals relevant to the Sun and the solar wind data. NICT is one of the contributing institutionsfor real-time data receiving networks. Because information on the solar wind reaching the earthcan be acquired beforehand using STEREO data (assuming that time-dependent changes in thesolar wind structure is negligible), the solar wind and interplanetary magnetic field data fromACE and STEREO were compared. In terms of the solar wind velocity, a good correlation wasconfirmed between ACE and STEREO data despite a widening elongation. For the north-southcomponent of the interplanetary magnetic field, however, the correlation diminished rapidly witha widening elongation. These findings revealed that the interplanetary magnetic field has timeand space scales of several hours and several million kilometers, respectively. These findingsalso suggest that it is necessary to consider the uncertainties of north-south magnetic field com-ponent in using STEREO data for predicting geospace disturbances.

KeywordsThe Sun, The solar wind, STEREO, Real time beacon, Preceding monitoring

42 Journal of the National Institute of Information and Communications Technology Vol.56 Nos.1-4 2009

Sun to the earth is needed to implement quan-titative space weather forecasting on a mid- tolong-term basis.

NASA’s ACE (Advanced CompositionExplorer) was launched at the L1 point in thesolar-terrestrial system in 1997. Although themain mission of the spacecraft is for scientificresearch, this is the first interplanetary space-craft to support a function for continuouslybroadcasting solar wind and solar energeticparticles data in a real-time beacon mode foroperational space weather forecasting. The L1point in the solar-terrestrial system is locatedabout 1.5 million km from the earth towardthe Sun, and represents one of the best loca-tions for observing the Sun and the solar wind.Because of a very low bit rate, real-time bea-con signals can only carry low-resolution data,but the data being continuously broadcastedby those signals aids in continuous monitoringof the solar wind conditions. A framework forreceiving solar wind data around the world hasbeen built and commissioned into service bythe National Oceanic and AtmosphericAdministration (NOAA) and US Air Force(USAF) with help from NICT and the U.K.’sRutherford Appleton Laboratory (RAL) toapply such data to space weather forecast-ing［3］［4］. We have used this solar windmonitoring data obtained from ACE in imple-menting day-to-day forecasting activities, pro-viding space environment information ser-vices, conducting real-time MHD simulations,and more.

However, the L1 point - located only about1.5 million km (or 0.01 astronomical unit)upstream from the earth - only allows predic-tions of about one hour in advance. Extendingthe lead time of the prediction requires infor-mation on solar wind variations at furtherupstream region, which we call “inner helios-phere (interplanetary space between the Sunand the earth)”. In the past, however, the onlyavailable means of observing the inner helios-phere was limited, to the remote sensing of thesolar wind by observing interplanetary scintil-lation (IPS), the Sun, and the solar corona bythe ground-based or satellite observation,

resulting in an inadequate coverage of innerheliospheric observation.

On October 25, 2006, NASA’s STEREOwas launched on a scientific research missionto explore the Sun and the inner heliosphere.The two spacecrafts — STEREO-A andSTEREO-B — had the primary functions ofobserving the Sun and inner heliosphere in astereoscopic view, and measuring the in-situplasma environment. Although the major pur-pose of STEREO observation is for scientificresearch, the data obtained from these space-crafts is also useful for space weather forecast-ing. Therefore, NASA decided to equipp real-time space weather beacons on the STEREOexplorers to broadcast the Sun and the solarwind data observed by STEREO in real time,though at a low resolution.

Data is also being received from STEREOreal-time space weather beacons in coopera-tion with overseas research institutions andsimilar organizations. Though the coverage ofdata acquisition is not 100%, data has thus farbeen received with reasonable success. AtNICT, an antenna facility for VLBI experi-ment is used to receive such data［1］. Asidefrom being used in day-to-day forecastingactivities, the receiving data is reviewed andanalyzed to explore new forecasting applica-tions. We examine the correlation betweenvariations in the solar wind as observed by theSTEREO and ACE using the solar wind data,and analyze the characteristic stability of thatdata for actual monitoring of the solar wind inadvance.

2 Preceding monitoring of thesolar wind using STEREO-B

Since the period of the Sun’s revolution isabout 27 days to the earth, the variations ofthe solar wind parameters are known to show27-day recurrence (Figure 1). Since STEREO-B is located forward from the earth (ACE) rel-ative to the solar revolution, the solar windparameters obtained from STEREO-B areexpected to arrive at the earth (ACE) after thetime lag tlag passed, assuming that the solar

43NAGATSUMA Tsutomu et al.

difference of the distance in the radial direc-tion, elongation, and the time lag of solar windas estimated by Equation (1). The motions inheliographic latitude are marked by a red line(for STEREO-A), black line (for ACE), andblue line (for STEREO-B). Due to an inclina-tion of 7.15 degrees between the axis of solarrotation and that of the earth’s revolution, theheliographic latitude at the base of the solarwind observed by ACE will vary in amplitudeof ±7.15 degrees. This essentially holds truewith STEREO-A and -B as well. One shouldtherefore keep in mind that one of the possiblecauses of the difference observed in the solarwind or magnetic field variations betweenACE and STEREO is the difference in helio-graphic latitude. Further, because STEREO-A,ACE, and STEREO-B are located closest tothe Sun in this order, both the differencebetween STEREO-A and ACE and the differ-ence of the distance in the radial directionbetween ACE and STEREO-B are negative.As for elongation from the earth, bothSTEREO-A and STEREO-B travel away byabout 20 degrees a year. Time lags are foundto widen with increases in elongation. A solidline designates a time lag for a solar windvelocity of 400 km/s, and a dotted line desig-nates that for a solar wind velocity of 800

wind has a stationary structure (free fromtime-dependent changes). The following equa-tion is used to estimate the time lag tlag:

tlag=θ/（ωsun－Ωearth）－（ωsun /（ωsun－Ωearth））・（（l 1－l 0）/Vsw）…………………………（1）

where, ωsun denotes the angular velocity ofthe Sun’s rotation, Ωearth the angular velocityof the earth’s revolution, l1 and l0 is the dis-tance from the Sun to STEREO-B and thatfrom the Sun to the earth, respectively, andVsw is the solar wind velocity observed bySTEREO-B. The orbit of STEREO-B some-what outside the earth’s orbit and graduallymoves away (rearward) from the direction ofthe earth’s orbital motion. Conversely, Theorbit of STEREO-A somewhat inside theearth’s orbit and gradually moves away towardthe direction of the earth’s orbital motion. Ifthe distance between the Sun and the earth,that between the Sun and STEREO-A, andthat between the Sun and STEREO-B wereequal, the time lag would be determined solelyby elongation. However, since there is a dif-ference in each distance in reality, informationabout the solar wind velocity is necessary toestimate the time lag.

Figure 2 shows (from top to bottom)orbital motions in heliographic latitude, the

Fig.1 The sector structure of the solarwind corotation driven by the solarrotation

Fig.2 Orbital trajectories of ACE, and ofSTEREO (top) heliographic latitude,(middle top) distance difference inthe radial direction, (middle bot-tom) elongation, and (bottom) esti-mated time lag

44 Journal of the National Institute of Information and Communications Technology Vol.56 Nos.1-4 2009

km/s. As shown in the diagram, the higher thesolar wind velocity, the wider the time lag.

According to Fig. 2, the time lag ofSTEREO-B reached a little less than two daysfrom about the end of 2007 to the middle of2008, suggesting that if the solar wind do havea stationary structure, the variations of thesolar wind can be predicted (through preced-ing monitoring) about two days ahead byusing data from STEREO-B.

3 Comparing solar wind varia-tions observed by ACE andSTEREO

To examine the feasibility for precedingmonitoring of the solar wind based onSTEREO data, variations in solar windobserved by ACE and that by STEREO werecompared. Figure 3 plots the hourly averagesof solar wind variations (in velocity, densityand temperature) as observed by ACE and bySTEREO during December 2007 without timelag correction. In this study, we used level 2data. Two fast solar winds flowing at about600 to 700 km/s were observed during theone-month period, indicating that fast solarwinds had been detected by STEREO-B,ACE, and STEREO-A in this order.

Figure 4 similarly plots solar wind varia-tions observed during December 2007, where

the times of solar wind variations observed bySTEREO-A and STEREO-B were correctedusing time lags estimated by using Equation(1) above for comparison with ACE observa-tions. The solar wind variations observed byall three spacecrafts were found in exception-ally good agreement based on the time lagcorrection process.

For comparison, each observation of theinterplanetary magnetic field was plotted withtime lag correction as shown in Figure 5.From the top to bottom, the diagram repre-sents the magnetic field strength (BT), the R-component of the RTN (Radial TangentialNormal) coordinate system, the T-component,and the angle of the R-T plane [sector struc-ture] (0 degrees in the R-axis direction and 90degrees in the T-axis direction). In the dia-gram, the RTN coordinate system has its R-component points from the center of the Sunto a spacecraft, T-component points crossproduct of solar rotational axis and R-compo-nent, and lies in the solar equatorial plane(towards the west limb). N component pointsR×T in the right-hand side system. The corre-lation between magnetic field variations asobserved by each spacecraft tends to remainrelatively low compared with the solar windvelocity variations. This might be consideredas a superposition of MHD waves in thequasi-static interplanetary magnetic field

Fig.3 Solar wind variations observed byACE and STEREO during December2007 (without time lag correction)

Fig.4 Solar wind variations observed byACE and STEREO during December2007 (with time lag correction)

45NAGATSUMA Tsutomu et al.

structure. Moreover, the increased magneticfield intensity observed by STEREO-A aroundthe 353rd day of the year may be associatedwith magnetic field variations of ICME. Onthe other hand, a relatively good correlationwas found between sector structures observedby ACE and STEREO.

Figure 6 plots variations of the correlationcoefficient between two different solar windobservations for Carrington rotation periodsnumbered from #2055 to #2075. The upperpanel in the diagram represents variations inthe correlation coefficient of the solar windvelocity; the lower panel represents variationsin the correlation coefficient of the north-south component (Bz component) of the inter-planetary magnetic field. Each red circlepoints to a correlation coefficient betweenSTEREO-A and ACE, with each blue circlepointing to a correlation coefficient betweenACE and STEREO-B, and each black circle toa correlation coefficient between ACE and theACE data for the previous Carrington rotationperiod. As for the solar wind, high correlationsevidently exist as a whole, with correlationcoefficients of 0.6 or higher generally beingobserved between STEREO and ACE duringthe periods of observation, and thus are higherthan the correlation with ACE data for the pre-vious Carrington rotation period. However,

the correlation coefficient dropped about threeperiods as marked in yellow open circle.Time-dependent changes in the solar windstructure during this period might be one ofthe causes. As for variations of Bz, the corre-lation coefficient down to 0.5 or below aboutthree periods passing after Carrington rotationperiod #2055. This may reflect the status ofthe solar wind structure, because the Bz com-ponent is directed virtually perpendicularly tothe plane of the solar rotation. The wideningelongation of STEREO orbit with increases inthe Carrington period suggested that the mag-netic field structure and time-dependent varia-tions had typical time and space scales of sev-eral hours and several million kilometers,respectively.

4 Summary and conclusions

The correlations between the solar windvariations observed by STEREO and by ACEwere examined using the solar wind and themagnetic field data obtained from these space-crafts. As a result, the correlation on the solarwind velocity did not diminish noticeably as awhole, despite a widening elongation, exceptfor the period when time-dependent changesoccurred in the solar wind structure. This sug-gests that the space weather forecasting based

Fig.5 Interplanetary magnetic field varia-tions observed by ACE and STEREOduring December 2007 (with timelag correction)

Fig.6 Variations of the correlation coeffi-cient between two different solarwind observations for Carringtonrotation periods from #2055 to#2075Upper panel: Solar wind velocityLower panel: North-south component of the

interplanetary magnetic field

46 Journal of the National Institute of Information and Communications Technology Vol.56 Nos.1-4 2009

on STEREO-B is basically acceptable. As forthe Bz component, a key parameter from theperspective of geospace disturbances, the cor-relation diminished rapidly with a wideningelongation, down to 0.5 or below about threeperiods after Carrington #2055. These find-ings revealed that typical time and spatialscales for north-south component of interplan-etary magnetic field estimated to be severalhours and several million kilometers, respec-tively.Based on these results, information ofthe solar wind velocity and the sector structurefrom STEREO-B can be used as input para-meters of empirical models and numericalsimulation, but not information about Bz in itsraw form. For this reason, it is necessary tomodel the uncertainties of the Bz componentand then consider a probabilistic scheme offorecasting. In addition, such tasks as how toassess the impact of time-dependent variationsin ICME and the solar wind structure remainto be solved, but the solar wind data providedby the STEREO should prove useful in pre-dicting the status of the magnetosphere andgeospace disturbances several days inadvance.

We will develop the techniques for remov-ing telemetry noise, calibrating physical quan-tities, and handling lack of measurements inreal-time beacon data, and then perform spe-cific data processing, such as time lag correc-tion and coordinate transformation, by usingthe resultant solar wind prediction data to con-duct a magnetospheric global MHD simula-tio［2］and drive an empirical geomagneticactivity forecasting model［5］, in order to pre-dict in a quantitative manner the status of themagnetosphere and geospace disturbancesseveral days in advance.

Acknowledgements

The Caltech ACE Science Center(http://www.srl.caltech.edu/ACE/ASC) pro-vided the ACE data for this study. We wish toexpress our thanks to the MAG and SWEPAMresearch teams. The UCLA IGPP Stereo DataServer (http://aten.ipgg.ucla.edu/ssc/stereo)provided the STEREO data. We also wish tothank the IMPACT (magnetic field data) andPLASTIC (solar wind data) research teams.

References01 M. Akioka, Y. Kubo, T. Nagatsuma, and K. Ohtaka, “Monitoring of the Solar Activity and Solar Energetic

Particles”, Special issue of this NICT Journal, 2-1-1, 2009.

02 H. Shinagawa, “Significance and Importance of the Integrated Space Weather Simulation”, Special

issue of this NICT Journal, 2-3-1, 2009.

03 T. Maruyama, S. Watanabe, K. Ohtaka, and H. Shimazu, “Real time solar wind monitor using ACE satel-

lite”, Rev. Comm. Res. Lab., Vol. 43, pp. 285–290, Jun. 1997. (in Japanese)

04 E. Sagawa, S. Watanabe, K. Ohtaka, H. Shimazu, and R. D. Zwickl, “Real time data reception of ACE

and IMAGE satellites”, Vol. 49, pp. 55–67, 2002.

05 T. Nagatsuma, “Monitoring and Forecasting of Geospace Disturbances, and its Importance”, Special

issue of this NICT Journal, 2-2-1, 2009.

47NAGATSUMA Tsutomu et al.

AKIOKA Maki, Dr. Sci.

Senior Researcher, Project PromotionOffice, New Generation NetworkResearch Center

Solar Physics, Optical System, SpaceWeather

NAGATSUMA Tsutomu, Dr. Sci.

Research Manager, Space EnvironmentGroup, Applied ElectromagneticResearch Center

Solar-Terrestrial Physics

OHTAKA Kazuhiro

Research Manager, Network SecurityIncident Response Group, InformationSecurity Research Center

Space Weather

MIYAKE Wataru, Dr. Sci.

Professor, School of Engeneering,Department of Aeronautics and Astronautics, Tokai University

Space Environment