Coronal sources and in situ properties of the solar winds sampled by ACE during 1999-2008
Hui Fu, Bo Li, Xing Li, Zhenghua Huang, Chaozhou Mou, Fangran Jiao, Lidong Xia
aa r X i v : . [ a s t r o - ph . S R ] M a y Solar PhysicsDOI: 10.1007/ ••••• - ••• - ••• - •••• - • Coronal sources and in situ properties of the solarwinds sampled by ACE during 1999-2008
Hui Fu · Bo Li · Xing Li · Zhenghua Huang · Chaozhou Mou · Fangran Jiao · Lidong Xia ∗ c (cid:13) Springer ••••
Abstract
We identify the coronal sources of the solar winds sampled by the ACEspacecraft during 1999-2008, and examine the in situ solar wind properties as afunction of wind sources. The standard two-step mapping technique is adoptedto establish the photospheric footpoints of the magnetic flux tubes along whichthe ACE winds flow. The footpoints are then placed in the context of EIT 284 ˚Aimages and photospheric magnetograms, allowing us to categorize the sourcesinto four groups: coronal holes (CHs), active regions (ARs), the quiet Sun (QS),and “Undefined”. This practice also enables us to establish the response to solaractivity of the fractions occupied by each kind of solar winds, and of their speedsand O /O ratios measured in situ. We find that during the maximum phase,the majority of ACE winds originate from ARs. During the declining phase,CHs and ARs are equally important contributors to the ACE solar winds. TheQS contribution increases with decreasing solar activity, and maximizes in theminimum phase when QS appear to be the primary supplier of the ACE winds.With decreasing activity, the winds from all sources tend to become cooler, asrepresented by the increasingly low O /O ratios. On the other hand, duringeach activity phase, the AR winds tend to be the slowest and associated withthe highest O /O ratios, and the CH winds correspond to the other extreme,with the QS winds lying in between. Applying the same analysis method to theslow winds only, here defined as the winds with speeds lower than 500 km s − , wefind basically the same overall behavior, as far as the contributions of individualgroups of sources are concerned. This statistical study indicates that QS regionsare an important source of the solar wind during the minimum phase. Keywords:
Solar Wind, sources . Solar wind, properties . Solar Cycle Shandong Provincial Key Laboratory of OpticalAstronomy and Solar-Terrestrial Environment, Institute ofSpace Sciences, Shandong University, Weihai, 264209, China Department of Physics, Aberystwyth University,Ceredigion, Wales, UK, SY23 3BZ ∗ correspondence addressed to [email protected] SOLA: ver4.tex; 11 August 2018; 4:21; p. 1 u et al.
1. Introduction
Identifying the source regions of the solar wind is important both as a fundamen-tal issue in solar physics (Antiochos et al. , 2012) and from the space environmentperspective (e.g., Luhmann et al. , 2002, and references therein). This practicedates back to the era when the solar wind was first measured (Snyder andNeugebauer, 1966; Nolte and Roelof, 1973; Neugebauer et al. , 1998, see alsoPoletto 2013 for a historic overview). With the solar wind data accumulatedthroughout several solar activity cycles in both near-ecliptic and polar orbits,scenarios have emerged as to how the solar wind sources evolve with solar ac-tivity. This concerns not only the solar winds sampled by individual spacecraftbut also the solar winds throughout the heliosphere (Luhmann et al. , 2002).Traditionally, the studies on the solar wind sources start with categorizing thewinds into the fast (with proton speeds v over, say, 500 km s − ) and slow ones( v .
500 km s − ) (e.g., Schwenn, 2006). Regarding the fast solar wind (FSW),the coronal source is generally accepted to be coronal holes (e.g., Krieger, Timo-thy, and Roelof, 1973; Zirker, 1977; Gosling and Pizzo, 1999). Tracing the windsampled by Pioneer VI and
Vela , Krieger, Timothy, and Roelof (1973) were thefirst to associate the FSW with a coronal hole. Then Zirker (1977) suggested thatall coronal holes are sources of the FSW. Using the SOHO/SUMER data, theoutflows at the base of polar (Hassler et al. , 1999) and equatorial (Xia, Marsch,and Curdt, 2003) coronal holes were measured, with the results supporting thenotion that the FSW originates in coronal funnels (Tu et al. , 2005). On theother hand, while examining the ACE and
Ulysses data for four Carringtonrotations during the Cycle 23 maximum, Neugebauer et al. (2002) concludedthat a fraction of the FSW originate also from active regions.The sources of the slow solar wind (SSW) are substantially more complex.While there exists the consensus that the SSWs are associated with coronalstreamers, debates remain as to exactly where in or around streamers the SSWsoriginate. The scenario proposed for solar minimum conditions by Wang et al. (1998) suggests that there are two kinds of SSWs, with one originating fromstreamer stalks and the other from just inside coronal holes and immediatelyadjacent to streamers. The former source is consistent with the outmoving plas-moids found by SOHO/LASCO (Sheeley et al. , 1997; Wang et al. , 1998), whilethe latter source is corroborated by the SOHO/UVCS measurements (Abbo et al. , 2010) and consistent with the established inverse correlation of the flowtube expansion with the solar wind speed (Wang and Sheeley, 1990). However,even at solar minimum, this scenario remains to be complemented with theexpected source of the SSWs from inside streamers, either via direct flow ofthe plasma from the magnetically open fields in streamer cores (Noci et al. ,1997) or via the evaporation of plasmas from the magnetic arcades in streamerhelmets (Suess et al. , 1999, also Li et al. , 2005). Besides, using the methodof interplanetary scintillation (IPS) tomographic analysis, Kojima et al. (1999)found that yet another SSW source is the unipolar regions in the vicinity ofactive regions (ARs). A further and more direct study associating the SSW withARs comes with Hinode X-ray and EUV spectral observations, where the edgeof ARs was shown to host persistent upflows with speeds reaching 100 km s − SOLA: ver4.tex; 11 August 2018; 4:21; p. 2
CE solar wind sources (Harra et al. , 2008), which may account for up to 1/4 of the in situ SSW (Sakao et al. , 2007) provided that these upflows eventually turn into outflows. Indeed,van Driel-Gesztelyi et al. (2012) (also see Culhane et al. , 2014 and Mandrini et al. , 2014) showed that these upflows may access coronal magnetic fields thatopen into interplanetary space. In addition, using X-ray high temporal-spatialresolution images, Subramanian, Madjarska, and Doyle (2010) found that themagnetic reconnection of co-spatial open and closed magnetic field lines atcoronal hole boundaries creates the necessary conditions for plasmas to flowto large distances. This provides an explanation for largely-blue-shifted eventsobserved with EIS/Hinode (Madjarska et al. , 2012), indicating these plasmaoutflows are also a possible SSW source. Comparing the remote sensing andin situ measurements, Feldman, Landi, and Schwadron (2005) suggested thatthe SSW may also arise from the quiet Sun. When it comes to solar maximumconditions, SSWs are found to originate from small coronal holes and activeregions where open magnetic field lines exist (Neugebauer et al. , 2002; Wangand Sheeley, 2003; Liewer, Neugebauer, and Zurbuchen, 2004; Ko et al. , 2006;Schwenn, 2006; Wang, Ko, and Grappin, 2009)While the identified coronal sources of the solar wind are diverse, there seemto be an agreement on the approaches behind the identification procedure. First,unlike the solar wind speed itself, ionic charge states, especially those of oxygenand carbon, are suggested to be a telltale signature of the wind sources. Takeoxygen for example. The abundance ratio O /O measured in the in situsolar wind is generally accepted to reflect the electron temperature in the coronalsources, given that it does not vary with distance beyond a fraction of a solar ra-dius above the solar surface (Owocki, Holzer, and Hundhausen, 1983; B¨uergi andGeiss, 1986; Hefti et al. , 2000; Landi et al. , 2012b). Now that the temperaturesare different in different coronal regions, a comparison of the in situ charge statesthen allows one to associate the in situ wind with a particular coronal source(e.g., Zurbuchen et al. , 2000; Zurbuchen, 2001; Landi et al. , 2012a). With thisspirit, Zhao, Zurbuchen, and Fisk (2009) divided the non-transient solar windsinto two categories: those from coronal holes (CH winds) and those from outsidecoronal holes (non-CH winds) with O /O values lower and higher than 0.145,respectively. As a result, about 42% of the ecliptic solar wind was found to beof non-CH origin during 1998-2008. Second, a model of coronal magnetic field isoften indispensable. For this purpose, while sophisticated Magnetohydrodynamic(MHD) models are sometimes adopted (Abbo et al. , 2010), the potential-field-source-surface model (PFSS) and its variants have been in much wider use. Onthe one hand, this practice established the long-term trend of the wind speedbeing inversely correlated with the lateral expansion of the flow tubes (Wang andSheeley, 1990). On the other hand, applying the PFSS model with an archive ofthe synoptic magnetogram data leads Luhmann et al. (2002) to the distributionof sources of the heliospheric solar wind as a function of solar activity for nearlythree activity cycles. In particular, Luhmann et al. (2002) found that althoughpolar coronal holes exist for more than 80% of a solar cycle, they contributeto the ecliptic solar winds significantly only during half of a cycle. During theother half of a cycle, the near-ecliptic winds originate from mid- and low-latitudesources instead. SOLA: ver4.tex; 11 August 2018; 4:21; p. 3 u et al.
Given the diversity of the wind sources and the complexity of the activity-dependence of these sources during a solar activity cycle, the present study isintended to examine, in a statistical manner, the fractions taken up by the insitu solar winds from various sources from the activity maximum to minimum inCycle 23. To this end, we start with the in situ wind speed measurements, andadopt the standard two-step mapping procedure (Neugebauer et al. , 1998, 2002;Liewer, Neugebauer, and Zurbuchen, 2004) to trace the winds to their footpointsat the solar surface. We then examine the corresponding coronal images recordedby SOHO/EIT as well as photospheric magnetograms, and ask the questionwhere the footpoints are located: are they located in a coronal hole (CH), anactive region (AR), or the quiet Sun (QS)? The solar winds are therefore groupedaccordingly, enabling us to address the question how their in situ properties differand evolve with different activity levels.Our study differs from previous studies with similar objectives or similarapproaches in the following aspects. First, the approach combining a footpointtracing method with the context of coronal images follows closely the one inLiewer, Neugebauer, and Zurbuchen (2004), which is in turn built on Neuge-bauer et al. (1998, 2002) where the imaging data were not used. However, whileNeugebauer et al. (1998) focused on the Cycle 22–23 minimum, and Liewer,Neugebauer, and Zurbuchen (2004, also Neugebauer et al. /O values are a primary discriminator.We note that, given the uncertainties in both approaches, the results of thisstudy are meant not to be contrasted with but rather to complement ZZF09,with the hope that new light can be shed on the sources of the near-ecliptic solarwinds. Third, while both using the PFSS model and being statistical in nature,our study differs from the one by Luhmann et al. (2002) in that we also employthe imaging as well as magnetogram data to classify the sources instead of usingthe locations relative to the equator as in Luhmann et al. (2002). Fourth, giventhe considerable interest in and the complexities associated with the sources ofthe slow solar wind, we will analyze the ACE solar winds in general, and examinethe slow ones in particular. In Section 2, we describe the data and our methodof analysis. The results are then given in Section 3. Section 4 summarizes thepresent study, ending with some concluding remarks.
2. Data and analysis
The two-step mapping procedure used in the present study closely follows theone in Neugebauer et al. (1998, 2002); Liewer, Neugebauer, and Zurbuchen(2004). To initiate the procedure, we use daily averages of the solar wind speed
SOLA: ver4.tex; 11 August 2018; 4:21; p. 4
CE solar wind sources made with the
Solar Wind Electron, Proton, and Alpha Monitor (SWEPAM,McComas et al. , 1998) on board the
Advanced Composition Explorer (ACE,Stone et al. , 1998). Also used are the daily averages of the abundance ratiosO /O recorded by the Solar Wind Ion Composition Spectrometer (SWICS,Gloeckler et al. , 1998), and the magnetic field measurements with the
MagneticField Experiment (MAG, Smith et al. ,1998). Given that we are interested in thenon-transient solar winds, one immediate purpose for using the O /O ratiosis to eliminate from the ACE dataset those intervals occupied by interplanetary Coronal Mass Ejections (ICMEs). To do this, we adopt the same approach asin ZZF09 (see also Richardson and Cane, 2004) whereby we discard the data withO /O ratios exceeding 6 .
008 exp( − . v ), in which v is the wind speedin km s − . A detailed analysis by ZZF09 shows that this criterion adequatelyseparates ICMEs from the non-transient ambient winds, being reliable in 83.2%of the cases examined therein. The in situ data used in this study span the yearsbetween 1999 and 2008, hence encompassing nearly half of the Cycle 23.The mapping procedure involves two steps. First, the loci of the solar windsare found on the source surface, placed at a heliospheric distance of 2.5 R ⊙ asimplemented by the coronal magnetic field model. This is done via a ballisticapproach, whereby the longitude correction due to solar rotation is determinedby the time for a wind parcel to travel from the source surface to the spacecraft.Here a constant wind speed is used, and assumed to be the one measured byACE/SWEPAM. The wind parcel is then traced from the source surface to thephotosphere by following the magnetic field lines computed by using a PFSSmodel, provided in the PFSS package as part of the Solar Software. Insteadof using the synoptic magnetograms as was done in e.g., Neugebauer et al. (1998), this package uses, as the boundary data, the magnetograms measuredwith SOHO/MDI which are updated every 6 hours. It outputs the magnetic fieldvector on a 39 × ×
192 grid in spherical coordinates inside the source surface(for details, see Schrijver and De Rosa, 2003). It should be noted that as impliedby the mapping procedure, the magnetic polarity at the field line footpoint needsto be checked against the one measured in situ. Schrijver and De Rosa (2003)found that during 1997-2001, 83% of footpoint polarities matched the interplan-etary magnetic field (IMF) measurements at the Earth. In this work, we find asimilar behavior: the footpoint polarities are consistent with what is measuredby ACE/MAG in 81% of the data from 1999 to 2008. To ensure consistency, wedo not include in our further analysis those dates when the polarities at the twoends of the mapping procedure do not match. Table 1 presents, as a functionof time, the number of daily samples of the non-transient solar wind (secondcolumn, labeled “All sources”), which is sub-divided into the counts of the solarwinds from CHs (third column), ARs (fourth), QS (fifth), and Undefined sources(sixth). Given in the parentheses are the numbers that correspond to the caseswhere the magnetic polarities match. In total, during 1999-2008, 2124 samplesare examined in our further analysis, among which 615 (803, 425) samples areassociated with CHs (ARs, the QS). One can see that a significant mismatchtakes place in 2007 and 2008. This is possibly understandable given that, closeto the cycle minimum, the ACE spacecraft was close to the heliospheric currentsheet. When mapping the winds to the source surface, a small uncertainty may
SOLA: ver4.tex; 11 August 2018; 4:21; p. 5 u et al. lead to a wrong polarity. For future reference, Table 2 presents the comparisonbetween the footpoint polarity and the in situ one for the slow solar winds withspeeds lower than 500 km s − .The footpoints are then placed in the context of photospheric magnetogramsand the EUV images taken by the Extreme ultraviolet Imaging Telescope (EIT,Delaboudini`ere et al. , 1995) onboard SOHO (
Solar and Heliospheric Observa-tory , Domingo, Fleck, and Poland, 1995). While EIT operates at a number ofpassbands (Fe ix/x
171 ˚A, Fe xii
195 ˚A, Fe xv
284 ˚A and He ii
304 ˚A), wechoose the 284 ˚A one because the images recorded in this passband reflect thecorona at the highest altitude such that coronal holes are more visible. In thispassband, EIT takes full-Sun images with a pixel size of 2.6 ′′ four times a day.For consistency, the field line footpoints are compared with images taken ataround 13:00 UT on the day corrected for the wind travel time.The classification scheme is illustrated in Figure 1, where the EIT images(the left column) are overplotted with the footpoint locations represented by thered crosses. The photospheric magnetograms, on which our scheme also relies,are derived from the PFSS model and given in the right column. The scheme isdetailed as follows.A quantitative approach for identifying coronal hole boundaries is imple-mented, and the winds that have footpoints located within the hence identifiedcoronal holes (CHs) are classified as “CH winds” accordingly. This approach,which largely follows that in Krista and Gallagher (2009) and Ko et al. (2014),is illustrated in Figure 2. If a footpoint is located inside or close to an apparentlydark area, then a rectangular box (the white box in Figure 2a) is chosen to enclosethis part of the dark region and its surrounding area. An intensity histogramis constructed, and a multi-peak distribution is then obvious (Figure 2c). Thewell-defined minimum between the first two peaks then defines the thresholdfor identifying the CH boundary (see the contours in Figure 2a, also Figure 2b,which is the enlarged version of the part inside the box). On the one hand, thisscheme enables one to objectively define CH boundaries using the EUV imagesin only one passband; on the other hand, it is not influenced by the variation ofcoronal emissions with solar activity.A description of some technical details for implementing this scheme seemsnecessary. In practice, we started with asking the question whether there is adark region close to the traced-back footpoint. By “close”, we mean roughly“within 100 arcsecs”. If the answer is Yes, we then draw a rectangular box,varying in size but typically a few hundred arcsecs across, which encloses botha substantial part of the dark region and its surrounding area. The footpoint isalways within this box. It turns out as long as the box is sufficiently large, itssize does not significantly influence what one identifies as CH boundaries, forthe minima in the different histograms pertinent to different box sizes do notdiffer substantially. If the answer is NO, we visually choose the dark area thatis the closest to the footpoint, and use the same approach to delineate the CHboundary (see e.g., Figure 1d1). If there is no large apparent EUV CH altogether,then we use the threshold found for some obvious CH one or a few days priorto this particular day (An example is shown in Figure 1b1, where the CHs nearthe two poles are contaminated so significantly that the minimum between the SOLA: ver4.tex; 11 August 2018; 4:21; p. 6
CE solar wind sources first two peaks in the intensity histogram can hardly be discerned). The boxis substantially smaller than the disc size, we nonetheless use the threshold todelineate CH boundaries throughout the entire solar disc. A space-dependentthreshold may be more accurate for mapping CH boundaries on the entire disc,but our approach suffices given that our purpose is to examine whether thefootpoint is located inside a CH. Besides, as illustrated by Figure 2a, while asingle threshold is adopted, the contours outside the box (the dotted lines) alsooutline CHs rather accurately.Our association of a footpoint with an Active Region (AR) or the quiet Sun(QS) relies on the magnetic morphology of the photospheric regions embodyingthe identified footpoints. The most obvious features on the photosphere aremagnetic clusters, which are tentatively named “magnetically concentrated area(MCA)”. Intuitively speaking, MCAs correspond to strong magnetic fields. Tomake this definition more objective, the absolute value of the radial componentof the photospheric magnetic field | B r, ⊙ | computed from the PFSS model isused. We experimented with different contour levels, | B r, ⊙ | B , used for outliningMCAs presented in Figure 1 and the attached movie. In practice, if | B r, ⊙ | B is assigned a value 1 . − | B r, ⊙ | , then MCAs become welldefined. That the MCA morphology is not sensitive to some given | B r, ⊙ | B , aslong as it is in the mentioned range, suggests that MCAs have sharp boundaries.This is understandable considering that MCAs have a strong spatial gradientin | B r, ⊙ | . The thus-defined MCAs encompass all the active regions numberedby NOAA as provided by solarmonitor . However, not all MCA patches cor-respond to a numbered AR. Many of these turn out to correspond to plageswith magnetic field weaker than concurrent numbered ARs (see section 3.2in Zharkova et al. | B r, ⊙ | , respectively. In the MAX and DEC phases thethreshold is about 10-20 Gs, which is close to the lower bound (15 Gs) adoptedby Wang, Ko, and Grappin (2009) to identify slow solar winds from ARs.With CH boundaries quantitatively defined, when identifying AR and QSsources we need only to concern about the regions outside CHs. An AR sourceis defined when a footpoint is located inside an MCA that is a numbered ARby NOAA. Likewise, QS sources are defined when a footpoint is located outsideany MCA.With the present grouping scheme, what is unclassified is then named “Un-defined”, and corresponds to the case where a footpoint is inside some MCAthat is not numbered by NOAA. These sources may be associated with a decay-ing/developing AR, but it is also possible that they are distinct from AR sources(see, e.g., 06/25-06/27 2005 in the movie where the source is likely a QS one).This is why the word “Undefined” is chosen.Such a scheme will not overestimate the counts in the respective groups. First,the counts of AR and QS sources are not overestimated, since some footpointsdeemed “Undefined” may in fact be AR and QS sources. Second, the counts of http://solarmonitor.org SOLA: ver4.tex; 11 August 2018; 4:21; p. 7 u et al.
CH winds are not overestimated either, for the current definition of CHs excludesa fraction of CHs with overlying bright emissions. In any case, the counts in theUndefined group account for only a minor fraction of the samples (11.2%, 9.5%,14.5%, 10.4%, 9.3%, 18.5%, 17.9%, 19.5%, 18.0%, and 5.3% for the years 1999to 2008, respectively).Before proceeding, several remarks on our approach seem in order. The firstremark is on the reliability of the PFSS model, given its apparently oversim-plification of imposing a spherical source surface, neglecting volumetric electriccurrents between the source surface and the photosphere, and supposing purelyradially directed field lines outside the source surface. Nonetheless, a detailedcomparison study by Riley et al. (2006) demonstrated that the magnetic fieldconfiguration computed by the simple PFSS model agrees well with the onefound in sophisticated MHD computations, provided that both models are drivenby the same line-of-sight magnetograms. From the practical point of view, themagnetically open regions obtained by the PFSS model well match the coronalhole regions identified in, say, the He I 10830 synotpic diagrams (Levine, 1982;Neugebauer et al. , 1998, 2002; Schrijver and De Rosa, 2003). A good way to makesure that the traced-back footpoints are reasonably accurate is to compare thecurrent SolarSoft PFSS results with some other calculations. To address this, werandomly chose three Carrington Rotations in the MAX, DEC, and MIN phases,and compared our derived footpoints with those derived from the PFSS modelwhere the magnetogram input is from the Wilcox Solar Observatory (WSO).We found that the fraction of the days when the two different sets of footpointsbelong to the same open field region is 80.7% for CR 1969, 84% for CR 2005,and 82.6% for CR 2054. Nevertheless, let us stress that the fraction that the twosets do agree is substantial enough that the statistical study we conduct can bedeemed reliable.Another source of uncertainty may come from the mapping procedure, par-ticularly in view of the simple ballistic treatment involved in the first step. Asdemonstrated by Nolte and Roelof (1973) (also see Neugebauer et al. , 2002),while the solar wind may experience some acceleration beyond the source sur-face, this effect may be counter-balanced by the near-Sun corotation. Actually afurther evidence lending us confidence with this mapping procedure is that, wheninspecting the footpoints on a consecutive basis (please see online animation1 attached to Figure 1), one can see an orderly distribution of the locationsof footpoints. They stay in a particular group for several days before movingto another group. In addition, assuming that the uncertainty with the sourcelongitude at the source surface is ± ◦ (Nolte and Roelof, 1973), we selecttwo Carrington rotations in each sub-interval (see Figure 3a) and examine howwell our classification scheme works. This is done by tracing the photosphericfootpoint from a locus on the source surface 10 ◦ eastward or westward of thenominal locus, and then examining whether the footpoint is located in a differentarea in the EIT images. We found that at maximum activity, about 30% of thefootpoints indeed are associated with an area different from what we identifiedusing the nominal locus. During the declining and minimum phases, however,this mismatch reduces to .
20% and 10% of the cases examined, respectively.
SOLA: ver4.tex; 11 August 2018; 4:21; p. 8
CE solar wind sources
3. Results
Having categorized the winds, we then address the question of the percentageeach kind of wind occupies, and how this evolves in response to solar activity.Figure 3a presents the monthly average of the smoothed sunspot numbers from1999 to 2008. Three sub-intervals, labeled MAX, DEC and MIN, are then definedaccording to the level of solar activity. Figure 3b presents the percentage of theCH (shaded blue), AR (red), QS (green) and Undefined (orange) winds in thisperiod. These percentages are yearly averaged values, and add up to unity ineach year. One can see that in the sub-interval MAX, the QS supplies only asmall fraction of the winds sampled by ACE ( ∼ ∼ − . v into 6 bins uniformly spaced between 200 and 800 km s − , group theO /O ratios into 6 bins uniformly spaced between 0.0 and 0.6, then presentin Figure 4 a contour plot in the v -O /O space the counts of the winds fromdifferent sources as labeled. The left, middle, and right columns correspond tothe intervals MAX, DEC, and MIN, respectively. Consider the interval MAX(left column) first. One notices that the majority of the winds correspondsto an O /O ratio larger than 0.145, which we recall is the criterion thatZZF09 employed to separate CH winds from non-CH ones. However, a more SOLA: ver4.tex; 11 August 2018; 4:21; p. 9 u et al. detailed analysis like ours indicates that not all winds that have O /O ratioslower than the nominal value of 0 .
145 are from CHs. Conversely, winds withO /O exceeding 0 .
145 are not necessarily non-CH ones. Now consider theyears 2007 and 2008, labeled MIN. One can see from Figure 4 (right column)that the O /O ratios tend to be low, with the majority being lower than0.145, meaning that if categorizing the ACE winds by this threshold, one wouldfind that nearly all the winds are from CHs. However, combining the footpointtracing approach with the EUV and magnetic field data, we find that the QSis the primary contributor to the ACE winds during this period. Furthermore,comparing Figures 4c1 with 4c3, one can see that the QS winds are distinct fromthe CH winds in that they tend to be substantially slower. To select the propersubset of the fast solar wind sampled by ACE that comes from CHs, it wouldbe almost unmistakable to choose those with speeds higher than 600 km s − and O /O lower than, say, 0 .
05. The contamination from the QS windswould be at most marginal, and that from the AR winds would be minimal.We note in passing that this practice has been successfully employed by Zhaoand Landi (2014). Regarding the declining phase (middle column of Figure 4),one finds that the possibility of distinguishing between CH winds and non-CHwinds lies in between the extremes of maximum and minimum conditions. Thisis particularly true in the speed dimension. The CH winds tend to be faster thanthe non-CH ones (mainly from ARs in this case), and the difference between thetwo tends to be more obvious than for the MAX phase, but appears significantlyless obvious than for the MIN phase.The O /O ratios for the CH winds during the MAX phase (Figure 4a1)require some explanation. There appears to be a fraction of the CH winds forwhich the O /O values exceed 0.26. If assuming ionization equilibrium, thiswould correspond to a freeze-in temperature exceeding 1.58 MK (Mazzotta et al. ,1998). This is beyond the currently accepted electron temperatures derived fromremote sensing measurements for CHs below 1.6 R ⊙ (Habbal, Esser, and Arndt,1993; Esser and Edgar, 2000 and references therein). This apparent discrepancyis not too worrisome given that this fraction of the CH winds tends to origi-nate from the boundaries between CHs and ARs, while the measurements madeby Habbal, Esser, and Arndt (1993) pertain to the region well inside a polar CH.Furthermore, as proposed by Esser and Edgar (2000), the electron distributionfunction may rapidly develop a non-Maxwellian character within the first severalsolar radii that eventually forms what is measured in situ as the halo electrons(Marsch, 2006). It is worth noting that this non-Maxwellian character is alsopossible to develop in AR and QS winds.The differences in the in situ properties of the winds from different sourcesare further examined in Figure 5, where (a) the wind speed and (b) the oxygencharge state ratio are plotted as a function of time. Given by the green, red,and blue curves are the parameters of the QS, AR, and CH winds, respectively.The standard deviations are given by the error bars for the corresponding values,which are slightly displaced from one another for display purposes. An immediateimpression from Figure 5 is that the CH winds tend to be the fastest, whilethe AR and QS winds have almost the equal speeds. And the O /O ratiosare lowest (largest) for CH (AR) winds, with the QS winds lying in between. SOLA: ver4.tex; 11 August 2018; 4:21; p. 10
CE solar wind sources
However, the considerable overlap in either the speed or O /O ranges meansthat neither of the two parameters, on its own, seems to suffice to discriminatethe wind sources. Regarding the activity-dependence of the parameters, one cansee from Figure 5a that the wind speeds in the three categories show a similarnon-monotonic behavior. Take the CH winds for instance. Their speed start fromrelatively low values ( ∼
500 km s − ) around MAX, rise to some 590 km s − in2003 before decreasing to around 500 km s − in 2004-2006, and then graduallyincrease to some 550 km s − toward the MIN phase. Moving on to Figure 5b, onecan see that the overall tendency of O /O ratios in response to solar activityis opposite to that of the speed, as would be expected given the well-establishedinverse correlation of the two parameters (Wang and Sheeley, 2003; Wang, Ko,and Grappin, 2009). Nonetheless, one can see that the O /O ratios fromdifferent groups of winds differ more significantly than the speeds do: Note themarked difference in the O /O values in the CH winds from those in the ARwinds in the whole period. It is noteworthy that with decreasing activity, theO /O ratios in all three types of solar winds tend to decrease, which agreeswith Lepri, Landi, and Zurbuchen (2013). The O /O values in AR (CH)winds are 0.26 (0.16) during MAX, and decrease to 0.10 (0.05) during MIN. Given the considerable interest in understanding the origins of the slow solarwind (SSW), it is informative to apply the same practice to the slow windsalone. In the present study, a SSW is defined to be the wind with speeds lowerthan 500 km s − . Figure 6 examines the time evolution of the fractions of theSSWs coming from various sources during 1999-2008. Overall, the impressionin the MAX and DEC phases is similar to what one finds in Figure 3 wherethe solar winds as a whole were considered. This similarity to Figure 3b is notsurprising given that, as shown in Figure 4, most of the solar winds is on thelower side when the speed is concerned. During the MIN phase, the contributionfrom the QS to the slow wind is even more important than that to the overallsolar wind. This is also understandable in view of Figure 4c1, given that thesolar winds from CHs are largely fast ones.Figure 7 presents (a) the wind speeds and (b) the O /O ratios for the slowsolar wind as a function of time. As far as the wind speeds are concerned, one cansee that the speed in a given group does not show a systematic variation withsolar activity. In addition, there is no clear-cut difference in the speeds of thewinds from different groups. A stronger temporal variation and a more significantdifference in different groups of winds lie in the O /O values (Figure 7b).Overall, the O /O values for all the winds show a decrease with decreasingsolar activity, and they are substantially different for different groups. The dif-ferences in the O /O values in winds from different sources may be a resultof the intrinsic difference in the respective source properties, the magnetic fieldstrength being the most likely one. At any rate, this reinforces the notion raisedby Antiochos et al. (2012), who suggested that the compositional properties andtemporal variability serve better in differentiating the wind sources than thespeeds. SOLA: ver4.tex; 11 August 2018; 4:21; p. 11 u et al.
The SSW properties may be compared with previous studies. Wang, Ko, andGrappin (2009) suggested that the slow wind during 1998-2007 mainly containstwo components: one from small holes located in and around ARs with highO /O ratios during maximum, the other from the boundaries of large CHswith intermediate O /O values. Our approach suggests that the majorityof the former component indeed comes from ARs during maximum. However,the latter component may actually come from all the three kinds of sources (seeFigure 4).
4. Conclusion
The main purpose of this work is to examine, in a statistical sense, the sourcesof the solar wind sampled by ACE during 1999-2008 in general, and those ofthe slow solar wind in particular. To this end, we start with the in situ windspeed, and find the photospheric footpoints of the wind parcels by employing thestandard two-step mapping procedure (Neugebauer et al. , 1998, 2002) where thePotential Field Source Surface (PFSS) model (Schatten, Wilcox, and Ness, 1969;Altschuler and Newkirk, 1969) is used. We then associate the footpoints withvarious areas in the EUV images recorded by EIT in its 284 ˚A passband andphotospheric magnetograms. With this association we classify the ACE windsinto three groups: coronal hole (CH), active region (AR), and quite Sun (QS)winds. Our main results can be summarized as follows.i) During Cycle 23 maximum (years 2000 and 2001), ARs are the main con-tributor to the ACE winds, the contribution of CHs (QS) is ∼
20% (13%).The winds in this interval tend to be slow, and the AR winds correspond tosubstantially higher O /O values than the CH winds. During the decliningphase, the contributions from CHs and ARs both amount to roughly onethird. Overall, the fraction of the QS winds in this period is 17%, and tendsto increase with decreasing activity, accounting for 31% of the winds in 2006.During the Cycle 23–24 minimum (2007 and 2008), the contribution of CHs(ARs) is about 31% (15%), while the QS contribution is ∼ /O ratios. While both lowerthan CH winds, the speeds of AR and QS winds do not show a substantialdifference. A slightly more pronounced difference between AR and QS windsis seen in their O /O values, with AR winds tending to be slightly hotter.As for the dependence on solar activity of the winds from the same sources,overall with decreasing activity the winds tend to have lower O /O ratios.iii) The fractions occupied by the slow solar winds from different groups show adependence on solar activity similar to the case where solar winds from allspeed ranges are considered. This can also be said for the activity dependenceof O /O values. During the minimum phase, the QS contribution to theslow wind is even more important than its overall contribution, amounting to ∼ SOLA: ver4.tex; 11 August 2018; 4:21; p. 12
CE solar wind sources
Our results suggest that the quiet Sun is an important source of the ACEsolar winds around the cycle 23-24 minimum. A further study dedicated tothe examination of the properties of the source region and in situ properties ofthis particular QS wind is needed. Regarding the source regions, such propertiesas the magnetic field strength as well as magnetic topology will be of interest.Regarding the in situ properties, the abundances of low first-ionization-potential(FIP) elements relative to their photospheric values will be informative (Feld-man, Landi, and Schwadron, 2005; Wang, Ko, and Grappin, 2009). In addi-tion, it will be worthwhile looking for direct signatures of outflow in the QSby examining the Doppler shifts with the emission lines measured with eitherSOHO/SUMER (e.g., Xia, Marsch, and Curdt, 2003) or Hinode/EIS (e.g., Fu et al. , 2014; Brooks, Ugarte-Urra, and Warren, 2015).
Acknowledgments
We would like to thank the anonymous referee for helpful comments.We thank the ACE SWICS, SWEPAM, and MAG instrument teams and the ACE Science Cen-ter for providing the ACE data. SOHO is a project of international cooperation between ESAand NASA. Wilcox Solar Observatory data used in this study were obtained via the web site http://wso.stanford.edu courtesy of J.T. Hoeksema. The Wilcox Solar Observatory is currentlysupported by NASA. This research is supported by the China 973 program 2012CB825601,the National Natural Science Foundation of China (41174154, 41274176, and 41274178), andthe Ministry of Education of China (20110131110058 and NCET-11-0305). BL is also gratefulto the International Space Science Institute (ISSI) for providing the financial support to theinternational team on the origins of the slow solar wind.
References
Abbo, L., Antonucci, E., Miki´c, Z., Linker, J.A., Riley, P., Lionello, R.: 2010, Characterizationof the slow wind in the outer corona.
Advances in Space Research , 1400. DOI . ADS .Altschuler, M.D., Newkirk, G.: 1969, Magnetic Fields and the Structure of the Solar Corona.I: Methods of Calculating Coronal Fields.
Solar Phys. , 131. DOI . ADS .Antiochos, S.K., Linker, J.A., Lionello, R., Miki´c, Z., Titov, V., Zurbuchen, T.H.: 2012, TheStructure and Dynamics of the Corona–Heliosphere Connection.
Space Sci. Rev. , 169.
DOI . ADS .Brooks, D.H., Ugarte-Urra, I., Warren, H.P.: 2015, Full-Sun observations for identifying thesource of the slow solar wind.
Nature Communications , 5947. DOI . ADS .B¨uergi, A., Geiss, J.: 1986, Helium and minor ions in the corona and solar wind - Dynamicsand charge states.
Solar Phys. , 347.
DOI . ADS .Culhane, J.L., Brooks, D.H., van Driel-Gesztelyi, L., D´emoulin, P., Baker, D., DeRosa, M.L.,Mandrini, C.H., Zhao, L., Zurbuchen, T.H.: 2014, Tracking Solar Active Region OutflowPlasma from Its Source to the Near-Earth Environment.
Solar Phys. , 3799.
DOI . ADS .Delaboudini`ere, J.-P., Artzner, G.E., Brunaud, J., Gabriel, A.H., Hochedez, J.F., Millier, F.,Song, X.Y., Au, B., Dere, K.P., Howard, R.A., Kreplin, R., Michels, D.J., Moses, J.D.,Defise, J.M., Jamar, C., Rochus, P., Chauvineau, J.P., Marioge, J.P., Catura, R.C., Lemen,J.R., Shing, L., Stern, R.A., Gurman, J.B., Neupert, W.M., Maucherat, A., Clette, F.,Cugnon, P., van Dessel, E.L.: 1995, EIT: Extreme-Ultraviolet Imaging Telescope for theSOHO Mission.
Solar Phys. , 291.
DOI . ADS .Domingo, V., Fleck, B., Poland, A.I.: 1995, The SOHO Mission: an Overview.
Solar Phys. , 1.
DOI . ADS .Esser, R., Edgar, R.J.: 2000, Reconciling Spectroscopic Electron Temperature Measurementsin the Solar Corona with In Situ Charge State Observations.
Astrophys. J. Lett. , L71.
DOI . ADS .Feldman, U., Landi, E., Schwadron, N.A.: 2005, On the sources of fast and slow solar wind.
Journal of Geophysical Research (Space Physics) , 7109.
DOI . ADS . SOLA: ver4.tex; 11 August 2018; 4:21; p. 13 u et al.Fu, H., Xia, L., Li, B., Huang, Z., Jiao, F., Mou, C.: 2014, Measurements of Outflow Velocitiesin on-disk Plumes from EIS/Hinode Observations.
Astrophys. J. , 109.
DOI . ADS .Gloeckler, G., Cain, J., Ipavich, F.M., Tums, E.O., Bedini, P., Fisk, L.A., Zurbuchen, T.H.,Bochsler, P., Fischer, J., Wimmer-Schweingruber, R.F., Geiss, J., Kallenbach, R.: 1998,Investigation of the composition of solar and interstellar matter using solar wind and pickupion measurements with SWICS and SWIMS on the ACE spacecraft.
Space Sci. Rev. ,497. DOI . ADS .Gosling, J.T., Pizzo, V.J.: 1999, Formation and Evolution of Corotating Interaction Regionsand their Three Dimensional Structure.
Space Sci. Rev. , 21. DOI . ADS .Habbal, S.R., Esser, R., Arndt, M.B.: 1993, How reliable are coronal hole temperatures deducedfrom observations?
Astrophys. J. , 435.
DOI . ADS .Harra, L.K., Sakao, T., Mandrini, C.H., Hara, H., Imada, S., Young, P.R., van Driel-Gesztelyi,L., Baker, D.: 2008, Outflows at the Edges of Active Regions: Contribution to Solar WindFormation?
Astrophys. J. Lett. , L147.
DOI . ADS .Hassler, D.M., Dammasch, I.E., Lemaire, P., Brekke, P., Curdt, W., Mason, H.E., Vial, J.-C.,Wilhelm, K.: 1999, Solar Wind Outflow and the Chromospheric Magnetic Network.
Science , 810.
DOI . ADS .Hefti, S., Gr¨unwaldt, H., Bochsler, P., Aellig, M.R.: 2000, Oxygen freeze-in temperaturesmeasured with SOHO/CELIAS/CTOF.
J. Geophys. Res. , 10527.
DOI . ADS .Ko, Y.-K., Raymond, J.C., Zurbuchen, T.H., Riley, P., Raines, J.M., Strachan, L.: 2006,Abundance Variation at the Vicinity of an Active Region and the Coronal Origin of theSlow Solar Wind.
Astrophys. J. , 1275.
DOI . ADS .Ko, Y.-K., Muglach, K., Wang, Y.-M., Young, P.R., Lepri, S.T.: 2014, Temporal Evolution ofSolar Wind Ion Composition and their Source Coronal Holes during the Declining Phase ofCycle 23. I. Low-latitude Extension of Polar Coronal Holes.
Astrophys. J. , 121.
DOI . ADS .Kojima, M., Fujiki, K., Ohmi, T., Tokumaru, M., Yokobe, A., Hakamada, K.: 1999, Low-speedsolar wind from the vicinity of solar active regions.
J. Geophys. Res. , 16993.
DOI . ADS .Krieger, A.S., Timothy, A.F., Roelof, E.C.: 1973, A Coronal Hole and Its Identification as theSource of a High Velocity Solar Wind Stream.
Solar Phys. , 505. DOI . ADS .Krista, L.D., Gallagher, P.T.: 2009, Automated Coronal Hole Detection Using Local IntensityThresholding Techniques.
Solar Phys. , 87.
DOI . ADS .Landi, E., Alexander, R.L., Gruesbeck, J.R., Gilbert, J.A., Lepri, S.T., Manchester, W.B.,Zurbuchen, T.H.: 2012a, Carbon Ionization Stages as a Diagnostic of the Solar Wind.
Astrophys. J. , 100.
DOI . ADS .Landi, E., Gruesbeck, J.R., Lepri, S.T., Zurbuchen, T.H.: 2012b, New Solar Wind DiagnosticUsing Both in Situ and Spectroscopic Measurements.
Astrophys. J. , 159.
DOI . ADS .Lepri, S.T., Landi, E., Zurbuchen, T.H.: 2013, Solar Wind Heavy Ions over Solar Cycle 23:ACE/SWICS Measurements.
Astrophys. J. , 94.
DOI . ADS .Levine, R.H.: 1982, Open magnetic fields and the solar cycle. I - Photospheric sources of openmagnetic flux.
Solar Phys. , 203. DOI . ADS .Li, B., Habbal, S.R., Li, X., Mountford, C.: 2005, Effect of the latitudinal distribution oftemperature at the coronal base on the interplanetary magnetic field configuration and thesolar wind flow.
Journal of Geophysical Research (Space Physics) , 12112.
DOI . ADS .Liewer, P.C., Neugebauer, M., Zurbuchen, T.: 2004, Characteristics of active-region sources ofsolar wind near solar maximum.
Solar Phys. , 209.
DOI . ADS .Luhmann, J.G., Li, Y., Arge, C.N., Gazis, P.R., Ulrich, R.: 2002, Solar cycle changes in coronalholes and space weather cycles.
Journal of Geophysical Research (Space Physics) , 1154.
DOI . ADS .Madjarska, M.S., Huang, Z., Doyle, J.G., Subramanian, S.: 2012, Coronal hole boundariesevolution at small scales. III. EIS and SUMER views.
Astron. Astrophys. , A67.
DOI . ADS .Mandrini, C.H., Nuevo, F.A., V´asquez, A.M., D´emoulin, P., van Driel-Gesztelyi, L., Baker, D.,Culhane, J.L., Cristiani, G.D., Pick, M.: 2014, How Can Active Region Plasma Escape intothe Solar Wind from Below a Closed Helmet Streamer?
Solar Phys. , 4151.
DOI . ADS .Marsch, E.: 2006, Kinetic Physics of the Solar Corona and Solar Wind.
Living Reviews inSolar Physics , 1. DOI . ADS .Mazzotta, P., Mazzitelli, G., Colafrancesco, S., Vittorio, N.: 1998, Ionization balance for opti-cally thin plasmas: Rate coefficients for all atoms and ions of the elements H to NI.
Astron.Astrophys. Suppl. , 403.
DOI . ADS . SOLA: ver4.tex; 11 August 2018; 4:21; p. 14
CE solar wind sourcesMcComas, D.J., Bame, S.J., Barker, P., Feldman, W.C., Phillips, J.L., Riley, P., Griffee,J.W.: 1998, Solar Wind Electron Proton Alpha Monitor (SWEPAM) for the AdvancedComposition Explorer.
Space Sci. Rev. , 563. DOI . ADS .Neugebauer, M., Forsyth, R.J., Galvin, A.B., Harvey, K.L., Hoeksema, J.T., Lazarus, A.J.,Lepping, R.P., Linker, J.A., Mikic, Z., Steinberg, J.T., von Steiger, R., Wang, Y.-M.,Wimmer-Schweingruber, R.F.: 1998, Spatial structure of the solar wind and comparisonswith solar data and models.
J. Geophys. Res. , 14587.
DOI . ADS .Neugebauer, M., Liewer, P.C., Smith, E.J., Skoug, R.M., Zurbuchen, T.H.: 2002, Sources ofthe solar wind at solar activity maximum.
Journal of Geophysical Research (Space Physics) , 1488.
DOI . ADS .Noci, G., Kohl, J.L., Antonucci, E., Tondello, G., Huber, M.C.E., Fineschi, S., Gardner, L.D.,Korendyke, C.M., Nicolosi, P., Romoli, M., Spadaro, D., Maccari, L., Raymond, J.C.,Siegmund, O.H.W., Benna, C., Ciaravella, A., Giordano, S., Michels, J., Modigliani, A.,Naletto, G., Panasyuk, A., Pernechele, C., Poletto, G., Smith, P.L., Strachan, L.: 1997, Thequiescent corona and slow solar wind. In: Wilson, A. (ed.)
Fifth SOHO Workshop: TheCorona and Solar Wind Near Minimum Activity , ESA Special Publication , 75.
ADS .Nolte, J.T., Roelof, E.C.: 1973, Large-Scale Structure of the Interplanetary Medium, I: HighCoronal Source Longitude of the Quiet-Time Solar Wind.
Solar Phys. , 241. DOI . ADS .Owocki, S.P., Holzer, T.E., Hundhausen, A.J.: 1983, The solar wind ionization state as acoronal temperature diagnostic.
Astrophys. J. , 354.
DOI . ADS .Poletto, G.: 2013, Sources of solar wind over the solar activity cycle.
Journal of AdvancedResearch , 215. DOI . ADS .Richardson, I.G., Cane, H.V.: 2004, Identification of interplanetary coronal mass ejectionsat 1 AU using multiple solar wind plasma composition anomalies.
Journal of GeophysicalResearch (Space Physics) , 9104.
DOI . ADS .Riley, P., Linker, J.A., Miki´c, Z., Lionello, R., Ledvina, S.A., Luhmann, J.G.: 2006, A Com-parison between Global Solar Magnetohydrodynamic and Potential Field Source SurfaceModel Results.
Astrophys. J. , 1510.
DOI . ADS .Sakao, T., Kano, R., Narukage, N., Kotoku, J., Bando, T., DeLuca, E.E., Lundquist, L.L.,Tsuneta, S., Harra, L.K., Katsukawa, Y., Kubo, M., Hara, H., Matsuzaki, K., Shimojo, M.,Bookbinder, J.A., Golub, L., Korreck, K.E., Su, Y., Shibasaki, K., Shimizu, T., Nakatani,I.: 2007, Continuous Plasma Outflows from the Edge of a Solar Active Region as a PossibleSource of Solar Wind.
Science . DOI . ADS .Schatten, K.H., Wilcox, J.M., Ness, N.F.: 1969, A model of interplanetary and coronal magneticfields.
Solar Phys. , 442. DOI . ADS .Schrijver, C.J., De Rosa, M.L.: 2003, Photospheric and heliospheric magnetic fields.
SolarPhys. , 165.
DOI . ADS .Schwenn, R.: 2006, Solar Wind Sources and Their Variations Over the Solar Cycle.
Space Sci.Rev. , 51.
DOI . ADS .Sheeley, N.R., Wang, Y.-M., Hawley, S.H., Brueckner, G.E., Dere, K.P., Howard, R.A.,Koomen, M.J., Korendyke, C.M., Michels, D.J., Paswaters, S.E., Socker, D.G., St. Cyr,O.C., Wang, D., Lamy, P.L., Llebaria, A., Schwenn, R., Simnett, G.M., Plunkett, S.,Biesecker, D.A.: 1997, Measurements of Flow Speeds in the Corona Between 2 and 30R.
Astrophys. J. , 472.
ADS .Smith, C.W., L’Heureux, J., Ness, N.F., Acu˜na, M.H., Burlaga, L.F., Scheifele, J.: 1998, TheACE Magnetic Fields Experiment.
Space Sci. Rev. , 613. DOI . ADS .Snyder, C.W., Neugebauer, M.: 1966, The Relation of Mariner-2 Plasma Data to SolarPhenomena. In: Mackin, R.J. Jr., Neugebauer, M. (eds.)
The Solar Wind , 25.
ADS .Stone, E.C., Frandsen, A.M., Mewaldt, R.A., Christian, E.R., Margolies, D., Ormes, J.F.,Snow, F.: 1998, The Advanced Composition Explorer.
Space Sci. Rev. , 1. DOI . ADS .Subramanian, S., Madjarska, M.S., Doyle, J.G.: 2010, Coronal hole boundaries evolution atsmall scales. II. XRT view. Can small-scale outflows at CHBs be a source of the slow solarwind.
Astron. Astrophys. , A50.
DOI . ADS .Suess, S.T., Wang, A.-H., Wu, S.T., Nerney, S.F.: 1999, Streamer Evaporation.
Space Sci.Rev. , 323. DOI . ADS .Tu, C.-Y., Zhou, C., Marsch, E., Xia, L.-D., Zhao, L., Wang, J.-X., Wilhelm, K.: 2005, SolarWind Origin in Coronal Funnels.
Science , 519.
DOI . ADS .van Driel-Gesztelyi, L., Culhane, J.L., Baker, D., D´emoulin, P., Mandrini, C.H., DeRosa,M.L., Rouillard, A.P., Opitz, A., Stenborg, G., Vourlidas, A., Brooks, D.H.: 2012, MagneticTopology of Active Regions and Coronal Holes: Implications for Coronal Outflows and theSolar Wind.
Solar Phys. , 237.
DOI . ADS . SOLA: ver4.tex; 11 August 2018; 4:21; p. 15 u et al.Wang, Y.-M., Sheeley, N.R. Jr.: 1990, Solar wind speed and coronal flux-tube expansion.
Astrophys. J. , 726.
DOI . ADS .Wang, Y.-M., Sheeley, N.R. Jr.: 2003, The Solar Wind and Its Magnetic Sources at SunspotMaximum.
Astrophys. J. , 818.
DOI . ADS .Wang, Y.-M., Ko, Y.-K., Grappin, R.: 2009, Slow Solar Wind from Open Regions with StrongLow-Coronal Heating.
Astrophys. J. , 760.
DOI . ADS .Wang, Y.-M., Sheeley, N.R. Jr., Walters, J.H., Brueckner, G.E., Howard, R.A., Michels, D.J.,Lamy, P.L., Schwenn, R., Simnett, G.M.: 1998, Origin of Streamer Material in the OuterCorona.
Astrophys. J. Lett. , L165.
DOI . ADS .Xia, L.D., Marsch, E., Curdt, W.: 2003, On the outflow in an equatorial coronal hole.
Astron.Astrophys. , L5.
DOI . ADS .Zhao, L., Landi, E.: 2014, Polar and Equatorial Coronal Hole Winds at Solar Minima: Fromthe Heliosphere to the Inner Corona.
Astrophys. J. , 110.
DOI . ADS .Zhao, L., Zurbuchen, T.H., Fisk, L.A.: 2009, Global distribution of the solar wind during solarcycle 23: ACE observations.
Geophys. Res. Lett. , 14104. DOI . ADS .Zharkova, V.V., Aboudarham, J., Zharkov, S., Ipson, S.S., Benkhalil, A.K., Fuller, N.: 2005,Solar Feature Catalogues In Egso.
Solar Phys. , 361.
DOI . ADS .Zirker, J.B.: 1977, Coronal holes and high-speed wind streams.
Reviews of Geophysics andSpace Physics , 257. DOI . ADS .Zurbuchen, T.H.: 2001, Heliospheric Magnetic Field Configuration and its Coronal Sources.In: Brekke, P., Fleck, B., Gurman, J.B. (eds.)
Recent Insights into the Physics of the Sunand Heliosphere: Highlights from SOHO and Other Space Missions , IAU Symposium ,585.
ADS .Zurbuchen, T.H., Hefti, S., Fisk, L.A., Gloeckler, G., Schwadron, N.A.: 2000, Magnetic struc-ture of the slow solar wind: Constraints from composition data.
J. Geophys. Res. ,18327.
DOI . ADS . SOLA: ver4.tex; 11 August 2018; 4:21; p. 16
CE solar wind sources
Table 1.
Number of daily solar wind samples analyzed in each year.Year All sources CH winds AR winds QS winds Undefined1999 237 (188) 32 (27) 148 (117) 28 (23) 29 (21)2000 261 (221) 50 (47) 145 (125) 36 (28) 30 (21)2001 272 (220) 43 (39) 155 (119) 38 (30) 36 (32)2002 289 (259) 65 (62) 152 (135) 38 (35) 34 (27)2003 277 (259) 142 (138) 71 (65) 34 (32) 30 (24)2004 242 (211) 63 (61) 94 (81) 34 (30) 51 (39)2005 250 (201) 78 (70) 70 (61) 56 (34) 46 (36)2006 262 (200) 60 (56) 55 (43) 94 (62) 53 (39)2007 257 (178) 68 (57) 36 (30) 114 (59) 39 (32)2008 259 (187) 61 (58) 29 (27) 157 (92) 12 (10)Sum 2606 (2124) 662 (615) 955 (803) 629 (425) 360 (281)
Table 2.
Number of daily slow solar wind samples analyzed in each year.Year All sources CH winds AR winds QS winds Undefined1999 188 (141) 20 (15) 119 (89) 23 (18) 26 (19)2000 204 (167) 26 (23) 123 (103) 29 (23) 26 (18)2001 245 (194) 37 (34) 139 (103) 37 (29) 32 (28)2002 238 (208) 40 (37) 135 (118) 32 (29) 31 (24)2003 111 (96) 35 (31) 36 (32) 20 (18) 20 (15)2004 194 (163) 36 (34) 83 (70) 28 (24) 47 (35)2005 172 (132) 34 (29) 52 (44) 47 (28) 39 (31)2006 205 (146) 38 (34) 40 (28) 81 (50) 46 (34)2007 187 (116) 38 (28) 30 (25) 91 (42) 28 (21)2008 169 (102) 25 (22) 15 (13) 122 (61) 7 (6)Sum 1913 (1465) 329 (287) 772 (625) 510 (322) 302 (231)
SOLA: ver4.tex; 11 August 2018; 4:21; p. 17 u et al.
Figure 1.
Illustration of the classification scheme of the ACE solar winds. The footpoints ofthe solar wind flow tubes are given by the red crosses, which are classified as being associatedwith a coronal hole (the first row), an active region (second), the quiet Sun (third) and someundefined source (bottom). The left column presents the EIT 284 ˚A images, while the rightcolumn gives the corresponding magnetic morphology of the photosphere. The green contoursoutline CH (left column) and Magnetically Concentrated Area (MCA) boundaries (right). Ananimation showing the sources during 1999–2008 is available online.
SOLA: ver4.tex; 11 August 2018; 4:21; p. 18
CE solar wind sources
EIT 284 20030611 (a) -200 -100 0 100 200X (arcsec)-300-200-1000100200 (b) (c) N ( I ) Figure 2.
The scheme for outlining coronal holes. Panel (a) presents the EIT 284 image on2003 June 11, when the traced-back footpoint (the red cross) is located close to a low-latitudeCH. The white box encloses the region for which the intensity histogram is constructed andpresented in panel (c), where the solid green line represents the minimum between the twopeaks, given by the two red dotted lines. This minimum is used as the threshold to delineate CHboundaries in (a), where the contours inside (outside) the box are given by the solid (dotted)lines. Panel (b) is an enlarged version of the part enclosed by the box in (a).
SOLA: ver4.tex; 11 August 2018; 4:21; p. 19 u et al. M on t h l y SS N MAX DEC MIN(a)99 00 01 02 03 04 05 06 07 08 09Year0.00.20.40.60.81.0 P e r c en t age UNQSARCH (b)
Figure 3.
Fractions of the ACE winds with different sources as a function of time. Panel(a) shows the temporal evolution of the smoothed monthly sunspot number during 1999-2008,which is further divided into the maximum (labeled MAX), declining (DEC) and minimum(MIN) phases. Panel (b) gives the percentage of the coronal hole (CH, blue), active region(AR, red), quiet Sun (QS, green) and undefined (UN, orange) winds.
SOLA: ver4.tex; 11 August 2018; 4:21; p. 20
CE solar wind sources
200 300 400 500 600 700 800Speed (km s -1 )0.00.10.20.30.40.50.6 O + / O + (a3) O + / O + (a2) O + / O + (a1)
200 300 400 500 600 700 800Speed (km s -1 ) (b3) (b2) (b1)
200 300 400 500 600 700 800Speed (km s -1 ) (c3) (c2) (c1) MAX DEC MIN QSARCH
Figure 4.
Dependence on solar activity of the distribution of solar winds from different sourcesin the speed–O /O space. The left (middle, right) column corresponds to the maximum(declining, minimum) phase, while the first (second, third) row represents the winds fromcoronal holes (active regions, the quiet Sun). Here the counts of solar wind samples in differentgroups are shown as contour plots with the contours equally spaced in each panel. SOLA: ver4.tex; 11 August 2018; 4:21; p. 21 u et al. S peed ( k m s - ) MAX DEC MIN(a)
99 00 01 02 03 04 05 06 07 08 090.00.10.20.30.4 O + / O + QSARCH (b)
Figure 5.
In situ properties of solar winds from different sources as a function of time during1999–2008. Here panels (a) and (b) are for the wind speeds and O /O ratios, respectively.The interval between 2000 and 2008 is further divided into three activity phases: maximum(MAX), declining (DEC) and minimum (MIN). The winds from coronal holes (CHs), activeregions (ARs) and the quiet Sun (QS) are represented by the blue, red, and green curves,respectively. As for the error bars, they represent the standard deviations in each year. SOLA: ver4.tex; 11 August 2018; 4:21; p. 22
CE solar wind sources
99 00 01 02 03 04 05 06 07 08 09Year0.00.20.40.60.8 P e r c en t age UNQSARCH
MAX DEC MIN
Figure 6.
Similar to Figure 3 but restricted to the slow wind with speeds less than 500km s − . SOLA: ver4.tex; 11 August 2018; 4:21; p. 23 u et al. S peed ( k m s - ) MAX DEC MIN(a)
99 00 01 02 03 04 05 06 07 08 090.00.10.20.30.4 O + / O + QSARCH (b)
Figure 7.