Chirality and Magnetic Configurations of Solar Filaments
aa r X i v : . [ a s t r o - ph . S R ] D ec Chirality and Magnetic Configurations of Solar Filaments
Y. Ouyang , , Y. H. Zhou , , P. F. Chen , , and C. Fang , School of Astronomy & Space Science, Nanjing University, Nanjing 210023, China; [email protected] School of Science, Linyi University, Linyi 276000, China Key Lab of Modern Astron. & Astrophys. (Ministry of Education), Nanjing University, China
ABSTRACT
It has been revealed that the magnetic topology in the solar atmosphere displays hemisphericpreference, i.e., negative/positive helicity in the northern/southern hemisphere, respectively.However, the strength of the hemispheric rule and its cyclic variation are controversial. Inthis paper, we apply a new method based on filament drainage to 571 erupting filaments from2010 May to 2015 December in order to determine the filament chirality and its hemisphericpreference. It is found that 91.6% of our sample of erupting filaments follow the hemisphericrule of helicity sign. It is found that the strength of the hemispheric preference of the quiescentfilaments decreases slightly from ∼
97% in the rising phase to ∼
85% in the declining phase of solarcycle 24, whereas the strength of the intermediate filaments keeps a high value around 96 ± ∼
63% to ∼
95% in the rising phase, and keeps a high value 82 ± Subject headings:
Sun: activity — Sun: magnetic fields — Sun: filaments, prominences
1. Introduction
The magnetic field in the solar atmosphere,which is responsible for solar eruptions (Chen2011), originates from the tachocline at the bot-tom of the convection zone, and rises up buoyantlyout of the solar surface. Because of the antisym-metry of the Coriolis force, the magnetic struc-tures in the solar atmosphere often display symme-try or antisymmetry between the northern and thesouthern hemispheres, i.e., the α -effect. A typicalexample is the Joy’s law (D’Silva & Choudhuri1993; Wang & Sheeley 1991; Fan et al. 1994). Onthe other hand, the interaction between the buoy-ant flux tubes and the turbulent convection zonemay weaken any hemispheric preference, i.e., theΣ-effect (Longcope et al. 1998a). In addition,there are other effects such as the surface motionsand magnetic diffusion. As a result, the strength of the hemispheric preference is expected to varyfor different proxies of the magnetic topology. Forexample, in terms of the current helicity, it wasfound that about 70 ±
12% of active regions followthe hemispheric rule (Seehafer 1990; Pevtsov et al.1995; Abramenko et al. 1997; Bao & Zhang 1998;Hagino & Sakurai 2005; Zhang 2006; Hao & Zhang2011; Gosain et al. 2013; Liu et al. 2014), i.e.,the helicity is negative in the northern hemi-sphere and positive in the southern hemisphere.In terms of the sunspot whorls, although ashigh as 80% of the H α whorls were found tobe counterclockwise (measured inwardly) in thenorthern hemisphere and clockwise in the south-ern hemisphere (Hale 1927; Richardson 1941;Ding et al. 1987). However, it was noticed thatonly a small fraction of sunspots show a vorticalstructure (Richardson 1941), and most sunspotwhorls include both clockwise and counterclock-1ise H α fibrils (Pvtsov et al. 2003). A similarcomplex situation happens to the X-ray sigmoids(Rust & Kumar 1996; Lim & Chae 2009), sincethey suffer from projection effects and even poten-tial magnetic fields may also have similar shapes.Another important proxy of the magnetictopology is the chirality of filament channels andtheir overlying coronal arcade, as proposed byMartin et al. (1992). Here the chirality of a fila-ment channel is defined to be dextral/sinistral ifthe axial magnetic field of the filament is to theright/left when viewed from the positive polar-ity side of the filament channel. The chirality offilament channels originates from several sources,e.g., the twist of the magnetic field before emerg-ing into the solar atmosphere, the interaction be-tween two neighboring flux systems, and solar sur-face motions (van Ballegooijen & Martens 1989;Zirker et al. 1997; van Ballegooijen et al. 1998;Mackay et al. 2000). These sources are further di-vided into 8 mechanisms, each of which producesthe chirality pattern consistent and/or inconsis-tent with the hemispheric rule (Yeates & Mackay2009). Therefore, it is important to determinethe percentage of solar filaments which follow thehemispheric rule. With a sample of 73 quies-cent filaments, they found that ∼
80% of themwere either dextral (with negative helicity) in thenorthern hemisphere or sinistral (with positive he-licity) in the southern hemisphere. In their paper,they also proposed a rule, called Martin’s rulehereafter, i.e., a dextral filament has right-bearingbarbs, whereas a sinistral filament has left-bearingbarbs. With such a one-to-one correspondence,one can immediately determine the chirality of afilament by looking at the H α image without thehelp of vector magnetograms. Applying this ruleto the H α images during 2000–2001, Pevtsov et al.(2003) confirmed that ∼ ±
3% of the quiescentfilament follow the hemispheric rule of helicity.Compared to quiescent filaments, active-region fil-aments seem to have a weaker hemispheric prefer-ence in helicity. For instance, Pevtsov et al. (2003)found that only ∼ ±
1% of active-region filamentsfollow the hemispheric rule. As an extreme result,Martin et al. (1994) stated that active-region fil-aments do not obey the hemispheric rule, withnearly half filaments being dextral or sinistral ineach hemisphere. One might argue that the con-tradictory results between Martin et al. (1994) and Pevtsov et al. (2003) are due to the smallsample, e.g., only 31 filaments in the former.However, later investigations with bigger sam-ples also led to opposing conclusions. For exam-ple, with 123 filaments, Yeates et al. (2007) foundthat 82% of all filaments follow the hemisphericrule, whereas Bernasconi et al. (2005) identifiedthe chirality of 658 filaments with an automateddetection method, and found that only 68% ofthem obey the hemispheric rule. Such a resultof weak preference might be due to the limitedresolution of the H α full-disk observations so thatthe automated method cannot work well. Whatis even worse, Martens et al. (2014) claimed thatthe hemispheric preference seems to disappear orreverse during parts of the declining phase.The discrepancy between different researchersmight be partly attributed to their chirality iden-tification method, where the filament chirality isdetermined by the bearing sense of the filamentbarbs according to Martin’s rule. The resultingstatistics is contaminated by three factors: (1)Projection effects: A left-bearing barb might befalsely identified to be right-bearing due to theprojection effects. This problem becomes moreserious as the filament is closer to the solar limb;(2) Limitation of Martin’s rule: As pointed outby Guo et al. (2010) and Chen et al. (2014), Mar-tin’s rule is applicable to the filaments supportedby a magnetic flux rope only. For filaments whichare supported by a magnetic sheared arcade, thecorrespondence between the filament chirality andthe bearing sense of the filament barbs would beopposite to the Martin’s rule; (3) Multi-sampling:The typical lifetime of a filament is weeks oreven months. Therefore, some filaments might becounted several times depending on their lifetime.In order to determine the strength of the hemi-spheric preference of the filament chirality, a bet-ter chirality identification method should be ap-plied. For example, Sheeley et al. (2013) proposedto use the “coronal cells” in the extreme ultra-violet (EUV) images of the Sun along with photo-spheric magnetograms to determine the chiralityof the filament channel. This method is applicablewhen the coronal cells are clear.Recently, Chen et al. (2014) proposed an indi-rect method to determine the chirality of a fila-ment without the help of vector magnetograms,which is independent of the magnetic type of2he filament, i.e., of normal-polarity or inverse-polarity. This method is based on the observa-tional fact that when a filament erupts, parts ofthe cold plasmas drain down along the two legs ofthe supporting magnetic field lines and impact thesolar surface, forming two conjugate draining sites(Zhou et al. 2006; Tripathi et al. 2013; Chen et al.2014). With respect to the magnetic polarity in-version line (PIL) or the filament spine, the conju-gate draining sites are either left-skewed or right-skewed, corresponding to dextral or sinistral chi-rality, respectively. One of the advantages of thismethod is that the draining sites are well sepa-rated and are close to the solar surface (i.e., notsuspended in the corona as the filament barbs),which do not suffer from the projection effects.Moreover, only the erupting stage of a filament isconsidered, by which a filament is sampled onlyonce.In this paper, we attempt to apply this newmethod to examine the strength of the hemi-spheric preference of filament chirality. The paperis organized as follows: The data sample and thechirality identification method are described in § §
3, which are discussedin § §
2. Data Sampling and Analysis
The
Solar Dynamics Observatory ( SDO ) mis-sion provides high-resolution EUV images andmagnetograms, which are observed by the At-mospheric Imaging Assembly (AIA, Lemen et al.2012) and the Helioseismic and Magnetic Imager(HMI, Scherrer et al. 2012), respectively. Sinceits launch in early 2010, the satellite has beenmonitoring the Sun continuously, covering boththe rising and the beginning of the decliningphases of solar cycle 24. Solar activities, includ-ing filament eruptions, were routinely recordedby the Heliophysics Event Knowledgebase (HEK,Hurlburt et al. 2012). From 2010 May 13 to 2015December 31, there are more than 1000 erupt-ing filaments/prominences. Roughly half of theseevents are prominences above or behind the limb(McCauley et al. 2015). With all these events ex-cluded, 576 filaments are found to erupt on thedisk. Among these events, only 5 erupting fil-aments have no clear draining sites. Therefore,a total of 571 erupting filaments are selected as our sample. The
SDO /AIA observes the Sun inseven EUV and three UV channels with a pixelsize of 0 . ′′ Global Oscillation Network Group ( GONG ) in H α (Harvey et al. 2011), where fila-ment barbs can be clearly recognized in many ofthem.When a filament erupts, generally two drainingsites are visible in EUV images, and the filamentchirality can be identified by the method proposedby Chen et al. (2014), i.e., the chirality of a fila-ment is dextral when the draining sites are left-skewed, or sinistral when the draining sites areright-skewed. The application of this method isillustrated in Figure 1 and explained as follows:The top panels are the sketch maps of the caseswith left skew of the draining sites (panel a, cor-responding to dextral chirality) and right skew ofthe draining sites (panel e, corresponding to sinis-tral chirality), respectively, where the dashed linesmark the magnetic PIL, the shaded areas repre-sent the filament spine, and the two circles in eachpanel mark the EUV brightenings associated withthe filament drainage. Panels (b–c) display theevolution of an erupting filament with dextral chi-rality, whereas panels (f–g) display the evolutionof an erupting filament with sinistral chirality. Toapply this method, we first check the SDO /AIAimages to obtain the locations of the brighteningsassociated with the filament draining, as markedby the circles in panels (c) and (g) of Figure 1. Inmany cases, the skew of the draining sites can thenbe determined. In some cases where the filamentis curved and the two draining sites are too closeto the magnetic PIL, we mark the locations of thedraining sites on the
SDO /HMI magnetogram, asshown by Figure 1(d, h). Here the magnetogramis derotated to the time of the EUV images. Withthe correspondence of the draining sites and themagnetic polarities, we can easily determine theskew of the draining sites, and hence the chiralityof the filament. It is noted that our chirality iden-tification method works well even when only onedraining site is visible (Bi et al. 2014).The validity of this chirality identificationmethod is confirmed with small samples by3uyang et al. (2015) and Hao et al. (2016) whofound that the results are in accordance withthe vector magnetograms, the skew of the twindimmings and the skew of flaring loops after fil-ament eruption. For a large sample like the onein this paper, we also checked the chilarity deter-mined by the skew of the associated flaring loopsin all the disk events, which was proposed byMartin & McAllister (1995). It is found that theresults obtained by the two methods are exactlythe same. It is also noted that the bright drainingsites of erupting filaments would not be confusedwith flaring patches since the brightening is pre-ceded by the dark draining filament materials,as indicated by Figures 3 and 7 in Ouyang et al.(2015). For all the events in our sample, we tracethe draining filament materials to determine thebright draining sites.
3. Results
With the method described in §
2, the chiral-ity of 571 erupting filaments is then determined.The top panel of Figure 2 plots the chirality dis-tribution of these filaments, where the horizontalaxis is the time and the vertical axis is the lati-tude. In this panel, the blue diamonds correspondto the filaments with dextral chirality (hence neg-ative helicity), and the red diamonds representthe filaments with sinistral chirality (hence pos-itive helicity). It is found that 307 out of 324, i.e.,94.8%, filaments in the northern hemisphere havenegative helicity, and 216 out of 247, i.e., 87.4%,filaments in the southern hemisphere have posi-tive helicity. Put together, 91.6% of our sample oferupting filaments follow the hemispheric rule ofhelicity sign.In comparison, we apply Martin’s rule to thesame sample, i.e., to examine the sign of helic-ity based on the bearing sense of the filamentbarbs, which are observed by
GONG H α tele-scopes. Among the 571 filaments, 7 events arenot visible in H α . Following Pevtsov et al. (2003)and Jing et al. (2004), for the remaining 564 fila-ments, the sign of helicity is assigned to be nega-tive/positive when the filament barbs are predom-inantly right-bearing/left-bearing. The resultingdistribution of the helicity sign is displayed in thebottom panel of Figure 2 with the same coordi-nates as the top panel. Similar to the top panel, the blue diamonds correspond to the filamentswith dextral chirality (hence negative helicity),and the red diamonds represent the filaments withsinistral chirality (hence positive helicity). Amongall the 564 H α filaments, it is found that 211 outof 322, i.e., 65.5%, of the filaments in the northernhemisphere have negative helicity, and 152 out of242, i.e., 62.8%, filaments in the southern hemi-sphere have positive helicity. It should be notedthat 76 filaments in the northern hemisphere and66 filaments in the southern hemisphere have nodiscernable barbs. If we exclude all the H α fila-ments without identifiable barbs, it is found that85.8% of the filaments in the northern hemispherehave negative helicity, and 86.4% of the filamentsin the southern hemisphere have positive helicity.As a whole, 86.0% of all the filaments with clearbarbs follow the hemispheric rule of helicity sign,which is 5.6% less than the value obtained by ourmethod using the filament draining sites.Following Engvold (1998), we divide the fila-ments in our sample into three types, i.e., (1)Quiescent ones, which are located in quiet re-gions with relatively weaker magnetic field, (2)Intermediate ones with one end in an active re-gion and the other in the quiet region, and (3)Active-region ones, which are located inside anactive region. It is found that among the 571filaments, there are 379 quiescent filaments, 100intermediate filaments, and 92 active-region fila-ments. Their time-latitude diagrams of the helic-ity sign are displayed in three rows of Figure 3,respectively, where the blue diamonds correspondto negative helicity, whereas the red diamonds topositive helicity. It is shown that the strength ofthe hemispheric rule is 93% for the quiescent fil-aments, 95% for the intermediate filaments, and83% for the active-region filaments, respectively.On the right side of each row, we sum up the cor-responding type of filaments with time and plottheir latitude distribution in histograms. It is re-vealed that each type of the filaments have a bi-modal distribution in latitude. Whereas the quies-cent filaments are distributed more broadly in lat-itude, the other two types are more concentratedin low latitudes. It is also seen that the quiescentand the intermediate filaments that are against thehemispheric rule are concentrated near the equa-tor, whereas the active-region filaments that areagainst the hemispheric rule are concentrated near4he latitude of ∼ ◦ in each hemisphere.
4. Discussions4.1. Strength of the Hemispheric Rule
It has been argued that the solar dynamoprocesses generate well-organized symmetric orantisymmetric magnetic patterns over the twohemispheres of the Sun (e.g., Zirker et al. 1997;Pevtsov et al. 2014). Important parameters char-acterizing the magnetic topology are the helic-ity and handedness or chirality, which can bequantified from vector magnetograms and imag-ing observations. As a result, a hemisphericrule was revealed, i.e., the helicity tends to benegative in the northern hemisphere and posi-tive in the southern hemisphere. However, thestrength of the hemispheric rule changes with dif-ferent proxies and even with different samples.As summarized by Wang (2013), the strengthlies in the range ∼ ∼ ∼ ∼ ∼ α images observed by the GONG network. It is found that there are no H α obser-vations for 7 events. For the remaining 564 fila-ments, 363, i.e., 64.3%, filaments follow the hemi-spheric rule. It should be noted that among the564 filaments, 142, i.e., 25.2%, of them have noclear barbs, hence their chirality cannot be deter-mined by the Martin’s rule. If we exclude these fil-aments, then 86.0% of all the filaments with clearbarbs follow the hemispheric rule. Such a value isabout 5.6% smaller than our estimate using fila-ment draining sites.Compared to some of the previous studies,our statistics shows a significantly stronger hemi-spheric preference of filament chirality, i.e., dex-tral in the northern hemisphere and sinistral in thesouthern hemisphere. For the quiescent filaments,Martin et al. (1994) analyzed 73 filaments andfound that the strength of the hemispheric rule is82%. Later, Pevtsov et al. (2003) analyzed 1436filaments, which leads to similar strength, i.e.,83%. However, our estimate is as high as 93%. Forthe intermediate filaments, Lim & Chae (2009)found that the strength of the hemispheric ruleis 84%. However, our estimate is up to 95%. For5he active-region filaments, whereas Pevtsov et al.(2003) estimated the strength of the hemisphericrule to be 76%, Martin et al. (1994) claimed thatthere is no hemispheric preference. However, wefound that 83% of the active-region filaments fol-low the hemispheric rule. The reason for the re-markable difference is probably that the previousauthors used the bearing sense of filament barbsto determine the chirality of the filaments, whichwould lead to misidentification of chirality whenthe filament is of the normal-polarity type, i.e., thecorresponding magnetic configuration is a shearedarcade. That is to say, we cannot judge the chi-rality of a filament by the bearing sense of itsbarbs (Chen et al. 2014; Ouyang et al. 2015), norcan we judge the chirality by the magnetic polari-ties of the two endpoints of a filament (Hao et al.2016). On the contrary, the chirality can be vi-sually determined by other patterns, such as theskew of the filament draining sites (Chen et al.2014), the skew of coronal loops or flaring loops(Martin 1998), and the skew of the twin dimmingsupon filament eruptions (Jiang et al. 2011). It was proposed that the hemispheric rulemight be time-dependent (e.g., Sakurai & Hagino2003; Choudhuri et al. 2004; Hao & Zhang 2011;Yang & Zhang 2012; Gosain et al. 2013). How-ever, the results were divergent. In terms of thecurrent helicity, the violation of the hemisphericrule was claimed to happen in any phase of a so-lar cycle. For example, Sakurai & Hagino (2003),Choudhuri et al. (2004), and Hagino & Sakurai(2005) claimed that the hemispheric rule might beopposite near solar minimum. However, Bao et al.(2000) found that the hemispheric rule is oppositeduring the rising phase of solar cycle 23. On thecontrary, Hao & Zhang (2011, 2012, 2013) sug-gested that the violation of the hemispheric rulehappens in the declining phase of solar cycle 23.In terms of the filament chirality, Mackay & van Ballegooijen(2001) theoretically predicted that the hemi-spheric rule may disappear during the decliningphase of a solar cycle. Martens et al. (2014) ap-plied an automated method to a big sample of fila-ments from 2001 to 2012. As a preliminary result,they showed that the hemispheric rule waxes andwanes. It is strongly present in 2001–2002 around solar maximum, but reverses in 2006–2007, whichwas approaching solar minimum. At other times,it is wholly absent. Again, they determined thefilament chirality by the Martin’s rule, which isvalid only for the filaments that are magneticallysupported by a flux rope according to Chen et al.(2014).In this paper, we used the filament drainingsites to determine the filament chirality. A quicklook at our results in Figure 3 immediately givesan impression that the hemispheric rule of thefilament chirality roughly holds well from 2010to 2015. In order to check the cyclic evolutionof the strength of the hemispheric rule, we cal-culate f , the percentage of the filaments whichfollow the hemispheric rule each year, and plotits evolution in Figure 4, where the top panelcompares the strength of the hemispheric ruleof all the filaments ( connected red squares ) withthe smoothed sunspot number (Clette et al. 2016),and the bottom panel compares the strength of thehemispheric rule among quiescent filaments ( redsquares ), intermediate filaments ( green squares ),and active-region filaments ( blue squares ). It canbe seen that as a whole, the filament chiralityfollows the hemispheric rule very well, and thestrength is ∼
90% in the rising phase of solar cy-cle 24, though slightly decreases to ∼
87% duringthe solar maximum and the declining phase. Af-ter dividing these filaments into three types, it isthen found that the strength of the quiescent fila-ments decreases slightly from ∼
97% in the risingphase to ∼
85% in the declining phase, whereasthe strength of the intermediate filaments keepsa high value around 96 ±
4% from 2010 to 2015.Only the active-region filaments show significantvariations. Their strength of the hemispheric rulerises from ∼
63% to ∼
95% in the rising phase, andkeeps a high value 82 ±
5% during the declin-ing phase. However, during a period from 2013June to 2014 January, which is around the solarmaximum, the hemispheric preference totally van-ishes. As seen from Figure 3, whereas no active-region filaments erupt in the northern hemisphereduring this period (marked by the two verticaldashed lines in the bottom panel of Figure 3),there are equal active-region filaments with dex-tral and sinistral chirality in the southern hemi-sphere. During this period, the sunspot area inthe southern hemisphere reaches its highest peak6n solar cycle 24, whereas the sunspot area inthe northern hemisphere reaches a local minimum(Deng et al. 2016).Similar to our result, Bao et al. (2001) alsofound that the fraction of active regions with re-versed helicity sign is higher near the solar max-imum of solar cycle 22. We are still not surewhether the absence of the hemispheric rule of theactive-region filaments (and hence the correspond-ing active regions, Wang et al. 2013) near solarmaximum happens every solar cycle. However, itis interesting to notice that active regions that donot follow the hemispheric rule are generally moreproductive in solar flares (Bao et al. 2001).
Magnetic measurements indicate that solar fil-aments can be divided into (1) inverse-polarity fil-aments, whose magnetic field component perpen-dicular to the magnetic PIL is opposite to whatexpected from the photospheric magnetograms,and (2) normal-polarity filaments, whose mag-netic orientation is the same as expected from thephotospheric magnetograms (Leroy et al. 1984;Bommier & Leroy 1998). Theoretically, the twotypes of filaments are described by the KR model(Kuperus & Raadu 1974) and the KS model(Kippenhahn & Schl¨uter 1957), respectively. Theformer corresponds to a flux rope, whereas thelatter corresponds to a sheared arcade. Theoriginal 2-dimensional models were later ex-tended to 3-dimensions (Antiochos et al. 1994;Aulanier & Demoulin 1998; van Ballegooijen 2004).With solar filaments being the progenitor of coro-nal mass ejections (CMEs, Chen 2011), it is nat-ural to think that the pre-eruption magneticstructure is a flux rope in some CME events,and a sheared arcade in other events (Gosling1999; Chen 2011; Cheng et al. 2014; Song et al.2014; Cheng et al. 2015), which was confirmed byOuyang et al. (2015) in observations. However, weare still lacking the information on the percentageof the two types of solar filaments.In 1980s, efforts were made to measure themagnetic field of solar prominences (Leroy et al.1984; Bommier & Leroy 1998). According toLeroy et al. (1984), ∼
25% of their >
900 mea-surements in a sample of 120 prominences corre-spond to the normal-polarity configuration. How- ever, as they discussed in their paper, such aratio is biased by several selection effects. Be-sides, there are different numbers of measure-ments for different prominences. With their re-sult, we still do not know how many prominenceshave the normal-polarity configuration. Unfortu-nately, there were no systematic measurementsof magnetic field for filaments/prominences in or-der to distinguish between inverse-polarity andnormal-polarity configurations. However, recentlyChen et al. (2014) proposed an indirect method todistinguish the two types of magnetic configura-tions, i.e., a dextral/sinistral filament with right-/left-bearing barbs respectively has an inverse-polarity configuration, whereas a dextral/sinistralfilament with left-/right-bearing barbs respec-tively has a normal-polarity configuration.This method can be summarized as follows:the filaments which obey Martin’s rule have theinverse-polarity configuration, whereas the fil-aments which disobey Martin’s rule have thenormal-polarity configuration. In order to checkthe percentage of each type of filaments, we ap-ply this method to our sample of 571 eruptingfilaments. Considering that 142 filaments haveno clear barbs, we scrutinize the high-resolution
SDO /AIA images to examine the bearing senseof the filament threads rather than the filamentbarbs, since filament threads are more reliablethan filament barbs in determining the bearingdirection (Martin et al. 2008). With the filamentthreads not identifiable in 7 filaments, we finallyget a sample of 564 filaments, including 372 quies-cent filaments, 100 intermediate filaments, and 92active-region filaments. It is found that Martin’srule and our new chirality identification methodagree with each other for 503 out of 564, i.e., 89%,filaments, i.e., these filaments are magneticallysupported by a flux rope, therefore are inverse-polarity filaments; However, Martin’s rule andour new method do not agree with each otherfor 61 out of 564, i.e., 11%, filaments, i.e., theyare magnetically supported by a sheared arcade,therefore are normal-polarity filaments. Amongthe 61 normal-polarity filaments, there are 15 qui-escent filaments, 9 intermediate filaments, and 37active-region filaments. In another word, amongour sample, 37 out of 92, i.e., 40%, active-regionfilaments are of the normal-polarity type, 9 outof 100, i.e., 9%, intermediate filaments are of7he normal-polarity type, and 15 out of 372, i.e.,4%, quiescent filaments are of the normal-polaritytype. These results are illustrated by the diagramsin Figure 5.It is noted here that in H α images, many fila-ments may have co-existing left-bearing and right-bearing barbs. Some are real, as discussed byGuo et al. (2010), others might be due to pro-jection effects. In this section, we used filamentthreads to check the bearing sense, instead. Itis found that only 6 filaments have co-existingleft-bearing and right-bearing barbs. We take thedominant sense for each filament. It is interest-ing to notice that the minority threads in eachfilament are generally located near one end of thespine, and remain intact upon eruption.
5. Summary
In this paper, we performed statistical analyseson the chirality and the magnetic configurations(inverse-polarity versus normal polarity) of the so-lar filaments which erupt on disk from 2010 May13 to 2015 December 31, covering both the risingphase and the beginning of the declining phases ofsolar cycle 24. The chirality is determined by anindirect method proposed by Chen et al. (2014),i.e., left-/right-skewed drainage corresponds tothe dextral/sinistral chirality, respectively. Thedetermination of the magnetic configuration isalso based on a method proposed by Chen et al.(2014), i.e., those filaments that follow the Mar-tin’s rule (Martin et al. 1994) are of the inverse-polarity type, and those that disobey Martin’srule are of the normal-polarity type. By study-ing a sample of 571 filaments, we obtained thefollowing results:(1) About 94.8% of the filaments in the north-ern hemisphere have negative helicity, and 87.4%of the filaments in the southern hemisphere havepositive helicity, indicating a significant hemi-spheric preference of helicity. As a whole, 91.6% ofour sample of erupting filaments follow the hemi-spheric rule of helicity sign. With the improvedmethod in determining the filament chirality, thestrength of the hemispheric rule is higher than pre-vious studies. It should be noted that the sta-tistical result is based on the erupting filaments.Those filaments which do not erupt during the diskpassage are not included in our sample. (2) Following the conventional way, we dividedthe filaments into three types, i.e., quiescent type,intermediate type, and active-region type. It isshown that the strength of the hemispheric rule is93% for the quiescent filaments, 95% for the inter-mediate filaments, and 83% for the active-regionfilaments, respectively.(3) Regarding the cyclic behavior of the hemi-spheric preference, it is found that the strengthof the quiescent filaments decreases slightly from ∼
97% in the rising phase to ∼
85% in the decliningphase, whereas the strength of the intermediatefilaments keeps a high value around 96 ±
4% allthe time. Only the active-region filaments showsignificant variations. Their strength of the hemi-spheric rule rises from ∼
63% to ∼
95% in the risingphase, and keeps a high value 82 ±
5% during thedeclining phase. However, during a half-year pe-riod around the solar maximum, the hemisphericpreference totally vanishes.(4) It is found that in our sample of erupting fil-aments, 89% are inverse-polarity filaments, whichare magnetically supported by a flux rope, whereas11% are normal-polarity filaments, which are mag-netically supported by a sheared arcade.The authors thank the referee for construc-tive suggestions and the
SDO , GONG , and theHeliophysics Events Knowledgebase (HEK) sys-tem teams for providing the data. This researchwas supported by the Chinese foundations NSFC(11533005 and 11025314) and Jiangsu 333 Project.
REFERENCES
Abramenko, V. I., Wang, T. J., & Yurchishin, V. B. 1997,Solar Physics, 174, 291Antiochos, S. K., Dahlburg, R. B., & Klimchuk, J. A. 1994,ApJ, 420, L41Aulanier, G., & Demoulin, P. 1998, A&A, 329, 1125Bao, S., & Zhang, H. 1998, ApJ, 496, L43Bao, S. D., Ai, G. X., & Zhang, H. Q. 2000, in IAU JointDiscussion, Vol. 7, IAU Joint DiscussionBao, S. D., Ai, G. X., & Zhang, H. Q. 2001, in IAU Sym-posium, Vol. 203, Recent Insights into the Physics ofthe Sun and Heliosphere: Highlights from SOHO andOther Space Missions, ed. P. Brekke, B. Fleck, & J. B.Gurman, 247Bernasconi, P. N., Rust, D. M., & Hakim, D. 2005,Sol. Phys., 228, 97Bi, Y., Jiang, Y., Yang, J., et al. 2014, ApJ, 790, 100 ommier, V., & Leroy, J. L. 1998, in Astronomical Societyof the Pacific Conference Series, Vol. 150, IAU Colloq.167: New Perspectives on Solar Prominences, ed. D. F.Webb, B. Schmieder, & D. M. Rust, 434Canfield, R. C., & Pevtsov, A. A. 1999,EOS Transactions, 80, 13Chen, P. F. 2011, Living Reviews in Solar Physics, 8Chen, P. F., Harra, L. K., & Fang, C. 2014, ApJ, 784, 50Cheng, X., Ding, M. D., & Fang, C. 2015, ApJ, 804, 82Cheng, X., Ding, M. D., Zhang, J., et al. 2014,ApJ, 789, L35Choudhuri, A. R., Chatterjee, P., & Nandy, D. 2004,ApJ, 615, L57Clette, F., Lef`evre, L., Cagnotti, M., Cortesi, S., & Bulling,A. 2016, Sol. Phys., arXiv:1507.07803 [astro-ph.SR] Deng, L. H., Xiang, Y. Y., Qu, Z. N., & An, J. M. 2016,AJ, 151, 70Ding, Y. J., Hong, Q. F., & Wang, H. Z. 1987,Sol. Phys., 107, 221D’Silva, S., & Choudhuri, A. R. 1993, A&A, 272, 621Engvold, O. 1998, in Astronomical Society of the PacificConference Series, Vol. 150, IAU Colloq. 167: NewPerspectives on Solar Prominences, ed. D. F. Webb,B. Schmieder, & D. M. Rust, 23Fan, Y., Fisher, G. H., & McClymont, A. N. 1994,ApJ, 436, 907Gosain, S., Pevtsov, A. A., Rudenko, G. V., & Anfinogen-tov, S. A. 2013, ApJ, 772, 52Gosling, J. T. 1999, Washington DC American GeophysicalUnion Geophysical Monograph Series, 111, 205Guo, Y., Schmieder, B., D´emoulin, P., et al. 2010,ApJ, 714, 343Hagino, M., & Sakurai, T. 2005, PASJ, 57, 481Hale, G. E. 1925, Proceedings of the National Academy of Science, 11, 691—. 1927, Nature, 119, 708Hao, J., & Zhang, M. 2011, ApJ, 733, L27Hao, J., & Zhang, M. 2012, in Astronomical Society of thePacific Conference Series, Vol. 456, Fifth Hinode ScienceMeeting, ed. L. Golub, I. De Moortel, & T. Shimizu, 69Hao, J., & Zhang, M. 2013,in IAU Symposium, Vol. 294, Solar and Astrophysical Dynamos and Magnetic Activity, ed. A. G. Kosovichev, E. de Gouveia Dal Pino, & Y. Yan,149Hao, Q., Guo, Y., Fang, C., Chen, P.-F., & Cao, W.-D.2016, RAA, 16, 001Harvey, J. W., Bolding, J., Clark, R., et al. 2011, in Bul-letin of the American Astronomical Society, Vol. 43,AAS/Solar Physics Division Abstracts heeley, Jr., N. R., Martin, S. F., Panasenco, O., & Warren,H. P. 2013, ApJ, 772, 88Song, H. Q., Zhang, J., Chen, Y., & Cheng, X. 2014,ApJ, 792, L40Tripathi, D., Reeves, K. K., Gibson, S. E., Srivastava, A.,& Joshi, N. C. 2013, ApJ, 778, 142van Ballegooijen, A. A. 2004, ApJ, 612, 519van Ballegooijen, A. A., Cartledge, N. P., & Priest, E. R.1998, ApJ, 501, 866van Ballegooijen, A. A., & Martens, P. C. H. 1989,ApJ, 343, 971Wang, Y.-M. 2013, ApJ, 775, L46Wang, Y.-M., & Sheeley, Jr., N. R. 1991, ApJ, 375, 761Wang, Y.-M., Sheeley, Jr., N. R., & Stenborg, G. 2013,ApJ, 770, 72Xu, H., Gao, Y., & Zhang, H. 2009, SCPMA, 52, 1749Yang, S., & Zhang, H. 2012, ApJ, 758, 61Yeates, A. R., & Mackay, D. H. 2009, Sol. Phys., 254, 77Yeates, A. R., Mackay, D. H., & van Ballegooijen, A. A.2007, Sol. Phys., 245, 87Zhang, M. 2006, ApJ, 646, L85Zhou, G. P., Wang, J. X., Zhang, J., et al. 2006,ApJ, 651, 1238Zirker, J. B., Martin, S. F., Harvey, K., & Gaizauskas, V.1997, Sol. Phys., 175, 27This 2-column preprint was prepared with the AAS L A TEXmacros v5.2. α barbs (hence the filament chirality cannot be determined) are labeled with green diamonds.12ig. 3.— The time-latitude distribution of chirality for the 379 quiescent filaments (panel a), 100 intermediatefilaments (panel c) and 92 active-region filaments (panel e), where the chirality is determined with thefilament drainage method, and the blue diamonds correspond to dextral chirality, whereas the red diamondsto sinistral chirality. On the right side of each row, the filaments are summed over time to get the latitudedistributions of the corresponding dextral and sinistral filaments.13ig. 4.— Top panel: Cyclic evolution of the sunspot number (black line) and f , the strength of the hemi-spheric rule obeyed by the filament chirality (connected blue squares). Bottom panel: Cyclic evolution of ff