Spatial Relationship between CMEs and Prominence Eruptions during SC 24 and SC 25

During their propagation, coronal mass ejections (CMEs) and prominences sometimes display a nonradial motion. During the years after the solar minimum, the CME central position angle tended to be offset closer to the equator compared to that of the associated prominence eruptions (PE). No such effect was observed during solar maximum. The purpose of this paper is to investigate the latitudinal offsets of CMEs with respect to their source regions. We study 256 events from SC 24 and SC 25, listed in the Coordinate Data Analysis Workshop Data Center. We analyzed the CMES radial offset from the associated PEs by comparing their latitudes in the plane of the sky. This work is an extension of the previous work by Gopalswamy et al., but with an independent data set. We have confirmed the systematic equatorward offset of CME from the solar source region for the rising phase of Solar Cycle 25. Our analysis of the relation between CME linear speed and PE-CME latitudinal offset indicated that the velocities of the deflected CMEs are mainly in the range of 200 and 800 km s−1. In this study, we compared the nonradial offsets for the rising and decay phases of SC 24 and our analysis has shown that during the decay phase more events deflected toward the pole can be observed. The observed variation is attributed to the presence of a substantial number of low-latitude coronal holes during the decay phase and to the influence from nearby active regions.


Introduction
The first reports on the nonradial motion of coronal mass ejections (CMEs) during their propagation in the solar corona are dated from the Skylab and Solar Maximum Mission (SMM) era (Hildner 1977;MacQueen et al. 1986).Hildner (1977) reported that the majority of the 93 CMEs studied during the Skylab mission (1973)(1974) tended to be deflected toward the equator.MacQueen et al. (1986), analyzing more than 70 CMEs during SMM, found similar results as Hildner (1977).
The offset of the leading edge of white-light CMEs from the near-surface activities such as prominence eruptions (Gopalswamy & Thompson 2000;Gopalswamy et al. 2000) and EUV eruptions (Plunkett et al. 2001;Sarkar et al. 2019) were reported.A comparison between prominence and CME trajectories indicated that both the prominence and the CME showed nonradial motions and therefore the CME deflected as a whole.
The statistical study by Plunkett et al. (2001) revealed that the CME's propagation in the corona is governed by the largescale solar magnetic field, which tends to push the CME material toward the equatorial plane.
Investigating the differences between the central position angle (CPA) of CMEs and their source-region position angle, Cremades & Bothmer (2004) found that the CMEs deflected to a lower latitude region by about 20°during the rise phase of solar cycle 23.The research points out that such nonradial motions would result in high-latitude CME propagating to the ecliptic plane and subsequently impacting the Earth.In fact, Gopalswamy et al. (2008) found that, even though CMEs originated at higher latitudes during the rise phase of solar cycle 23, the CMEs arrived at Earth as magnetic clouds indicating the deflection of the CMEs toward the ecliptic.Cremades et al. (2006) analyze the deflection of CMEs with respect to their source regions in the time period between 1996 and 2002 and report a systematic equatorward offset during the period of low solar activity.Gopalswamy et al. (2003), on the base of microwave observations from Nobeyama Radioheliograph, report a solarcycle dependence in the spatial relationship between the PEs and CMEs.The results of their statistical investigation, covering 200 PEs from the 23rd solar cycle, reveal a systematic offset to the equator of the associated CMEs in the rise phase, whereas no such offset was observed during solar maximum.Further, this result was confirmed for the rise phase of cycle 24 (Gopalswamy 2015).
The above discussion on CME deflection focused on the latitudinal deflection.It is now known that CMEs propagate nonradially away from nearby coronal holes (CHs) to areas of lower magnetic energy.Many studies suggest that two factors may be responsible for such CME nonradial motions, namely the influence of the background coronal magnetic field (MacQueen et al. 1986;Kilpua et al. 2009;Shen et al. 2011) and the fast solar-wind flow from polar coronal holes (Wang et al. 2004;Cremades et al. 2006).
Original content from this work may be used under the terms of the Creative Commons Attribution 4.0 licence.Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.Cremades et al. (2006) found a correlation between the deflection of CME position angles from their source region and the total area of coronal holes.Gopalswamy et al. (2009a) introduce the CH influence parameter (CHIP) depending on the area, average magnetic field, and distance from the CH on the disk to described its influence on CMEs propagation (see also Mohamed et al. 2012;Mäkelä et al. 2013).The authors point out that CMEs are deflected away from the coronal hole where the magnetic filed is stronger than the surrounding corona.Gopalswamy et al. (2010) considered a modified CHIP involving the square of the coronal hole magnetic field.
A quantitative analysis of CMEs trajectory through the corona made by Shen et al. (2011) confirmed that the deflection is mostly controlled by the background magnetic field.They found that magnetic field energy density influenced the early CME propagation and caused them to deflect toward the region with the lower magnetic energy density, e.g., the heliospheric current sheet (HCS).Further, this research was expanded by Gui et al. (2011), which analyzed the deflection in both latitude and longitude of 10 CMEs and confirmed that the deflection is mostly controlled by the background magnetic field and CME offset from radial propagation is toward the region of low magnetic energy density surrounding the HCS.Zuccarello et al. (2012) used 2.5D numerical MHD simulations and observations from the two STEREO spacecraft to analyze the CME latitudinal offset toward the HCS.They concluded that, during solar minima, the CMEs are easily deflected toward the equatorial plane and the deflection rate depends on the strength of both the large-scale coronal magnetic field (magnetic fields of the overlying helmet streamer) and the magnetic flux of the erupting filament.
The near-Sun deflection of some CMEs, besides the influence of coronal holes, was attributed to the magnetic pressure of active region fields in the eruption's vicinity or to the interaction between two CMEs.From long-wavelength radio and white-light observations, Gopalswamy et al. (2001) reported the detection of interaction between two coronal mass ejections and their deflection as a consequence of this interaction.The numerical study (Xiong et al. 2009) reports CME latitudinal deflection during CME-CME interaction.By combining remote sensing and in situ observations, Lugaz et al. (2012) analyzed the kinematic and evolution of two successive CMEs and conclude that, because of the interaction between them, they are deflected away from each other.
A number of studies have shown that the deflection may be caused by the strong magnetic fields in the active region (e.g., Sterling et al. 2011;Gopalswamy et al. 2014;Kay et al. 2015;Möstl et al. 2015;Wang et al. 2015).By analyzing the near-Sun deflection of a CME in connection to the magnetic field configuration in the source active region (AR), Wang et al. (2015) note that the CME was directed away from the AR within 2.5 R e .The work of Gopalswamy et al. (2014) and Möstl et al. (2015) on the CME from 2014 January 7 demonstrated that an observed sizable deflection (of about 37°) with respect to the source region was caused by nearby active region rather than CHs.Liewer et al. (2015) examined the 3D trajectories of five CMEs and found out that CME's nonradial propagation from the source was a result of magnetic pressure of active regions in the vicinity of the eruption.This resulted in CME asymmetric expansion, which was further followed by an apparent deflection.
From the above discussion, it is clear that CME deflection is an important phenomenon that has both science (the cause of deflection) and practical (CME arrival at Earth) implications.Therefore, the earlier statistical study on the offset between eruptive prominence and CME leading edge in cycle 23 by Gopalswamy et al. (2003) needs to be extended to the new solar cycle.In this paper, we follow up study to the work of Gopalswamy et al. (2003) and Gopalswamy (2015) but with a new data set.Instead of the Nobeyama radioheliograph data, we used EUV prominence eruption data obtained by the Solar Dynamics Observatory (SDO, Pesnell et al. 2012) to study the spatial relationship between CME and Prominence eruptions.In Section 2, we describe the data set used in this study.The main results are presented and discussed in Section 3. The conclusions are drawn in Section 4.

Data
We study the near-Sun position-angle (PA) offset of CMEs from the associated prominence eruptions (PEs) by comparing their latitudes.Our sample contains 256 events from solar cycle (SC) 24 and the beginning of SC 25.The latitudes were obtained from the PAs, which are measured counterclockwise from the solar North in degrees (Gopalswamy et al. 2003).The apparent latitude is computed as 90°-CPA for the eastern limb events and CPA-270°for western limb events.Here CPA refers to the central position angle (i.e., nose of the CME leading edge).We used the PA of the leading edge of the CME during its propagation in the plane of the sky, the so-called measurement PA (MPA).The PA offset between the initial PE location and the CME nose (in C3 when it is possible to measure) is considered.Then we converted the PAs to heliographic latitudes.In this study, the absolute values of the latitudes was used, since we are interested in the latitudinal offsets.
The initial PAs of the PEs are taken from the Catalog of Prominence Eruptions6 compiled from SDO's Atmospheric Imaging Assembly (AIA; Lemen et al. 2012) in the 304 Å passband (Yashiro et al. 2020).The catalog contains the PEs seen above the limb since May 2010 and provides the CPA, the apparent latitude, and the angular width of the eruptive prominences.
We also used the CME Catalogavailable online at the Coordinated Data Analysis Workshop (CDAW) Data Center (Yashiro et al. 2004;Gopalswamy et al. 2009b) and the measuring tool available therein.The catalog lists CMEs manually detected with the Large Angle and Spectrometric Coronagraph (LASCO; Brueckner et al. 1995) onboard the Solar and Heliospheric Observatory (SOHO; Domingo et al. 1995) and provides CME parameters such as CPA, linear, and second-order speeds, angular width, acceleration, mass, and kinetic energy.

Solar Cycle Variation of the Latitude Offset
Positional relationship between PEs and CMEs is solar-cycle dependent.During solar minimum periods, PEs start at higher latitudes and the corresponding CMEs appear at relatively lower latitudes.This indicates a systematic CMEs deflection toward the equator with respect to PEs was observed during periods of low solar activity.No such tendency was detected during the maximum activity, where the offsets are alike toward the pole and equator.
In Figure 1, we have presented an example event in which the CME and prominence propagation is shown as observed at the east solar limb on 2010 June 20.The prominence eruption started at PA of 70°at 23:34 UT the previous day.The initial PE direction is indicated by a yellow line in Figure 1(a).In the next 4 hr, the CME nose and the prominence underwent a significant equatorward deflection to PA of ∼97°.We also examined how the deflection changes with radial distance.In the Figure 2(b) we show the CME core and nose PA at various heliocentric distances.In that particular event, the nonradial motion stopped at about 9 R e .We performed a similar analysis for the 256 events.
For the majority of the events, the PA changed until the CME reached a certain height and then became stable.The bar plot in Figure 3(a) shows the distribution of all analyzed CMEs by the critical height above which the propagation transforms from nonradial to radial.We found that in 74% the maximum PA was attained between 5 R e and 7 R e , with the peak value occurring at around 6 solar radii.
We also examined how the critical height is dependent on the initial CME kinematics.The heights, in solar radii, at which the nonradial motion ceases are plotted against the linear speed of the CME in Figure 3(b).The linear regression to the data points is shown by the red line.It is clear that these two characteristics are related to one another.CMEs with linear velocities less than 900 km s −1 were found to exhibit nonradial motions up to 5-7 solar radii, while CMEs with higher velocities continued to exhibit nonradial motion upward to larger heights.Figure 4 shows the PE-CME latitude offset as a function of time for the period between 2010 and 2022.The PE PA is measured when prominence area reaches its maximum in the AIA images.The CMEs final position angle is calculated as an average value from points when deflection is no longer observed, with standard deviation ranging between 0.21 and 1.5.We distinguished the PEs, originating from the northern and southern hemispheres, by plotting them with different colors.The beginning of SC 25 (2019 December 1) is shown with a vertical dashed line.Positive offset values mean that the prominence eruption started at a higher latitude than the corresponding CMEs, and the observed deflection was toward the equator.Figure 4 shows that during the rising phases of SC 24 and SC 25 the offset had positive values, that means that the deflection was toward the equator.No such effect is observing during solar maximum period.The deflection solely to the equator for SC 24 ceases in 2011 December and after that time, the deflection toward the pole can also be observed.For the beginning of solar cycle 25, the observed nonradial offset is solely directed toward the equator.Hence, we confirmed the solar-cycle variations of latitude offsets during the rising phase of SC 25.
The average positive offset between the position angles of PEs and the associated CMEs for the rising phases seems to be similar between the two cycles, i.e., about 14°for SC 24 and about 15°for SC 25.Positive PE-CME offset is considered an indicator of the stronger polar field (Gopalswamy & Thompson 2000;Cremades et al. 2006;Panasenco et al. 2011).During the rise phase of the solar-cycle CHs are present in the polar regions.At the beginning of a cycle, the polar CHs are closer to active regions that produce the CMEs.Gopalswamy et al. (2012) compared the 22/23 and 23/24 minima properties and  found that the average PE-CME offset for the 23/24 minimum was 19°and during the 22/23 minimum it was 18°.Our analysis show that the average PE-CME offset for the 24/25 minimum was 16°.

Deflection and CME Kinematic Parameters
CME nonradial propagation is believed to be controlled by the background magnetic field along with certain CME kinematic parameters.Some recent studies (Wang et al. 2004;Xie et al. 2009;Shen et al. 2011;Sieyra et al. 2020) indicate that during solar minimum slower and wider CMEs tend to be deflected toward the equator due to the influence of strong magnetic field of polar coronal holes.It is considered that slow CMEs cannot penetrate through the magnetic field of the CH, but they can be guided by them away from CH.
Based of ForeCAT (Forecasting a CMEs Altered Trajectory) model, which predicts the CME deflection due to magnetic forces, Kay et al. (2015) analyzed the dependence of nonradial CME motions from its parameters, such as mass and velocity.The authors found that slower and massive CMEs can be deflected more easily than fast ones, and pointed out that the deflection is more sensitive to velocity than to mass.
That motivated us to analyze the relationship between CMEs propagation velocity, mass and deflection rate.In Figures 5(a),  (b) and (c) we present a scatterplot of latitudinal offset as a function of CME speed for rising, maximum and declining phases, respectively.For our study, we used the linear CME speed taken from the LASCO catalog in the CDAW database.The linear speed serves as the average speed within the LASCO FOV.The CME's mass varies from 10 13 to 10 16 g, with the majority of events having a mass about 10 15 g.CMEs with different mass are shown with different colors in Figures 5(a), (b), and (c).
The bar graph in Figure 5(d) shows the distribution of all analyzed CMEs by linear speed.The velocities of the deflected CMEs are distributed mainly between 200 and 800 km s −1 .About 76% of all the CMEs included in our study have speeds within this velocity range.It is evident from Figure 5(b) that the large number of events (about 45%) have velocities between 200 and 450 km s −1 .This result is consistent with the study of Xie et al. (2009), which as a result of statistical study of 27 coronal mass ejections, find that fast CMEs tend to deflect less than slow ones (450 km s −1 ).

Positional Relationship Between CMEs and PEs During
Solar Cycle 24: Rising versus Decay Phase We analyzed the CME offsets from their source-region EP during the rising and declining phases of SC 24 and compare the two phases.The PA offset between the initial PE location and the CME nose as a difference in latitude for the rising phase of SC 24 (2010-2012) is presented in Figure 6(a) as a histogram.
Positive values indicate that the PEs are at higher latitude than the CMEs (and hence the observed nonradial motions are toward the equator) and vice versa.The events with positive offsets are the most common.
For the rising phase of SC 24 we found that the PE-CME latitude was positive for 86% of the studied events and negative for 14% of them.This result confirms again the finding that during the periods of minimum solar activity, especially in the rising phase of the solar cycle, the offset between PEs and CMEs are predominantly positive indicating a deflection toward the equator.
We did the same analysis for the declining phase of SC 24 as well.To determine the duration of decay phase we used the SILSO (sunspot Index and Long-term Solar Observations) data of monthly mean total sunspot number, prepared by the Royal Observatory of Belgium, Brussels and online catalog.
We have defined the decay phase as the long period of decline in solar activity, from the end of 2014 November to 2019 December, when the new cycle 25 started.The result is presented in Figure 6(b).Mainly positive offset (equatorward deflection) was observed during the decay phase also.The results revealed that a significant number of events, namely 76.31% were deflected toward the equator, while 23.68% displayed poleward offsets.This is a surprising result because there are no strong polar coronal holes in the declining phase.It is important to note that compared to the ascending phase, a larger fraction of events undergo nonradial movements toward the pole in the decay phase.
In order to find the reason for the equatorward deflection in the decay phase, we have shown in Figure 7 the latitudinal distribution of PEs (green stars) along with the nearby coronal holes (red circles) during the declining phase of SC 24.The filled red circles are sized in accordance with the area of the CHs.The largest dimension refers to the area >20 km 2 .Some of the nearby active regions are presented with triangles in the figure.The CH area was determined using the Collection of Analysis Tools for Coronal Holes (CATCH) (Heinemann et al. 2019).Based on filtergrams taken by the AIA/SDO 193 Å, EIT/SOHO 195 Å, and EUVI/STEREO 195 Å CATCH allow the user to perform CH boundary detection, extraction, and analysis of their properties.In Figure 8 is shown an example for CH boundary detection for event, occurring on 2017 October 18.
According to Shen et al. (2011) and Gui et al. (2011) the deflection is caused by the variations in the background magnetic energy.The CMEs are deflected toward the region of the lowest magnetic energy, such as the HCS.Hence, the direction of the nonradial offset is related to the HCS location throughout the solar cycle, which is more inclined during the declining phase, compared to the rising.The latter can result in a broader range of deflections.
Another reason for the observed difference could be that, during the declining phase of the solar cycle, the magnetic field is more complex, in comparison to the rising phase.The direction of deflection can also be attributed to a local magnetic field (Kay et al. 2015).Hence, the deflection during the decay phase may be different because of the presence of large number of low-latitude CH, which can also affect the nonradial propagation toward the pole if appropriately located.Furthermore, the sources regions (PEs) are located at higher latitudes during the rising phase but located closer to the equator during the declining phase.Hence, the latitude of source region can be lower than that of the associated CMEs, which results in negative PE-CME latitudes.
The nonradial offset of the CMEs from the source region can be influenced by local magnetic field related to active regions in the vicinity of PE.Such cases are enclosed with black circles in Figure 7. From 38 studied events during the decay phase, 18 have CH deflection, and 20 of the observed offsets were attributed to the influence of the nearby ARs.Table 1 provides the exact numbers of events during the decay phase of SC 24, separated by the deflection reason (CH deflection or AR deflection) and direction, e.g., toward the equator or toward the poles.

Summary
We analyzed the near-Sun nonradial motions of CMEs with respect to the solar source.We examined 256 events from solar cycle 24 and the onset of SC 25 extending the previous work by Gopalswamy et al. (2003Gopalswamy et al. ( , 2012) ) and Gopalswamy (2015) with an independent data set.The results can be summarized as follows: 1.The equatorward offset of CME from the solar source region during its early stages, which was first identified during solar cycle 23 by Gopalswamy et al. (2012) and Gopalswamy (2015) for SC 24 has been confirmed for the rising phase of SC 25. 2. We found that the average positive offset for the rising phase of SC 24 was 14°.4,while for the rising phase of SC 25 this offset was 15°.Our analysis also showed that the average PECME offset for the 24/25 minimum was 16°, slightly smaller than of Gopalswamy et al. (2012), who reported an average offset of 19°for the 23/24 minimum, and during 18°for the 22/23 minimum.3. We analyzed the relation between CME linear speed and PE-CME latitudinal offset and found that the velocities of the deflected CMEs are mainly in the range of 200 and 800 km s −1 , with 45% of them between 200 and 450 km s −1 .4. During the rising phase of SC 24, 86% of the studied events had positive deflection and 14% had negative one, while for the decay phase of SC 24 76.31% have been deflected toward the equator and 23.68% showed poleward offsets.During minimum of solar activity the nonradial CME motions are mainly equatorward, but during the decay phase the percentage of the events deflected toward the pole is slightly higher.The observed variation is attributed to the presence of a substantial number of low-latitude coronal holes or ARs.We also found that more than half of the analyzed events during the decay phase have been deflected by the nearby ARs.
Figure 1(b) shows the CME PA, marked with a red line, at 02:36 UT.Finally, the CME was observed in C3 FOV at PA ∼97°at 04:20 UT and continued its propagation without deflection.In the last panel of Figure 1, the CME is shown at 03:26 UT, before it left the LASCO/C2 field of view.In Figure 1(c), the dashed red line points to the CME position angle at 03:26 UT, the red line marks the CME PA as measured at 02:26 UT, and the double dashed yellow line shows the direction of the initial PA of the PE.Thus the PA offset between the initial PE location and the CME nose was 27°the equator.The time variations of the CME core and nose are shown in Figure 2(a).

Figure 1 .
Figure 1.Deflection of the 2010 June 20 CME and PE.(a) AIA/SDO 304 Å image with outlined PE.The yellow line shows the PE CPA (b) SOHO/LASCO/C2 and AIA/SDO 193 Å composite image at 02:26 UT.The red line denotes the CME PA (c).SOHO/LASCO/C2 and AIA/SDO 193 Å composite image at 03:26 UT.The CME PA for 02:26 UT and 03:26 UT and PE PA are shown by red, dashed red and double dashed yellow lines, respectively.

Figure 2 .
Figure 2. (a) Variation of the position angle of CME core and nose as a function of time.Triangles, stars, and diamonds represent measurements, respectively in AIA FOV, from C2, and C3 coronagraphs of SOHO/LASCO.(b) CME Position angle (PA) at various heliocentric distances.The solid line is the fit to the data points, showing that the nonradial motion stopped at about 9 R e .

Figure 3 .
Figure 3. (a) The CME distribution by heliocentric distance at the transition point from nonradial to radial motion.(b) CMEs linear velocities versus heliocentric distance when the nonradial motion transforms into radial [solar radii].The red line is the linear regression to the data points.

Figure 4 .
Figure 4. Solar cycle variation of the latitude offset between PE and CME, as observed by LASCO.PEs originating by the northern/southern hemispheres are shown with red/green triangles.The vertical dashed line marks the beginning of SC 25.

Figure 5 .
Figure 5. PE-CME latitude offset as a function of CME speed: (a) for rising phases of SC 24 and SC 25 , (b) for maximum phase of SC 24, (c) for declining phase, (d) Distribution of all analyzed CMEs by linear speed.

Figure 6 .
Figure 6.(a) Positional relationship between CMEs and PEs as a difference in PA for rising phase of SC 24 .(b) Positional relation between CMEs and PEs as a difference in PA for decay phase of SC 24.

Figure 7 .
Figure 7. Latitudinal distribution of coronal holes and prominence eruptions during the declining phase of SC 24.CH, PE, and ARs positions are indicated by filled circles, stars, and triangles, respectively.Cases with possible deflection by AR are marked by black open circles.

Figure 8 .
Figure 8.An illustration of the CH boundary detection for the event that took place on the southeast solar limb on 2017 October 18.

Table 1
Deflected Events During the Decay Phase of SC 24