Summer Research Fellowship Programme of India's Science Academies

Associating CMEs to their source regions in solar cycle 23 & 24 and understanding their effects on different observational features of CMEs

Kavya Bindu M

Mangalore University, Mangalore 574199

Prof. Dipankar Banerjee

Indian Institute of Astrophysics, II block, Koramanagala, Bangalore 560034


Coronal Mass Ejections are huge eruption of plasma and magnetic field from the solar corona. The objective of the project is to study i) the fractional contributions of source regions of CMEs during different phases of solar cycle, ii) their latitudinal and longitudinal distributions, iii) latitude vs position angle plots to look for deflections of CMEs, iv) the distribution of CME width from different sources,v) dependency of CME width on position angle (if any),v) distribution of speed of fast and slow CMEs. The source regions of cme are categorised as, i) Active regions (AR) – Regions of strong magnetic field, visible as sunspot groups in the visible wavelength, ii) Active prominences (AP)- cme observed due to prominence eruption with the footpoints of the erupting prominence connected to ARs, iii) Prominence eruptions (PE)- cme due to prominence eruption from the quite regions. In this work all the three sources of CMEs are identified during different phases of solar cycle 23 & 24 by back propogating the CMEs onto the solar disc and their spatial locations are recorded in terms of latitude and longitude. The observations are made using Large Angle and spectroscopic Coronagraph (LASCO C2) onboard Solar and Heliospheric Observatory (SOHO) and the Atmospheric Imaging Assembly (AIA) onboard Solar Dynamic Observatory (SDO). The source regions on the back of the sun with respect to the Sun-Earth line are identified with the help of EUV 171 &304 onboard the twin spacecraft STEREO A,B. The results obtained in this project supplies useful information in understanding what fractional contributions of source regions from the active and quite Sun can we expect to CMEs during different phases of solar cycle and what impacts can they have on their kinematical features in the present and upcoming solar cycle.

Keywords: coronal mass ejection, source region of CME, solar cycle 23 &24


Coronal Mass Ejection (CME) is the large expulsion of magnetised plasma from the solar surface into the interplanetary space. In the white light coronographs they are observed as transient bright features expelling in the field of view (FOV). CMEs are massive structures weighing nearly 1011-1012 kg and emerging with the speed in the range 400 to 1000 km/s. CME involve many small sized eruptive phenomena like solar flares, filament/prominence eruption etc (1).

The CME observed in the solar corona may have the initiation due to the configurational changes in the photospheric magnetic field or due to pre-existing features such as filaments in the lower corona. (1)

The solar surface has regions of strong magnetic field called as Active regions which are seen as sunspots in visible wavelength.A CME may have an onset from the active regions due to i) magnetic reconnection or ii) small magnetic reconfiguration or iii) instabilities such as torous and kink instabilities. Such CME source regions are categorised as AR.(1)

Prominences are bright gaseous features which are anchored form the solar photosphere and extended into the outer corona. They are comprised of very dense, cool ionised gas. Prominences are often found in the quite regions, having relatively low magnetic field, on the sun. An eruption of a prominence from lower corona also constitutes CME categorised as PE.

A CME may occur due to eruption of a prominence which is embedded in the active region. Such CME are categorised as AP.

In this paper we have collected data for the rising phases of solar cycle 23(jan to may 1998 & march to june 1999) and for solar cycle 24 ( sept to dec 2010 & jan to june 2011). Rising phases is considered as the change in activity can be clearly observed as the cycle build up the activity from the solar minima.

For both the solar cycles studies are made for the rising phases so that comparison of the acquired data can be done for the solar cycles 23 & 24.

    Depiction of solar cycle 23 & 24.The region enclosed in the box is the raising phases for which the data is collected in this paper.


    The Coordinated Data Analysis Workshop (CDAW) catalog contains the CME which are manually identified by Large Angle and Spectroscopic COronograph (LASCO) onboard Solar and Heliospheric Observatory (SOHO) since 1996. The CDAWcatalog also contains the information of the associated parameters of CMEs such as angular width, position angle, linear speed etc, but has no information regarding the source locations of CMEs. (5)

    The catalog designed for the analysis of this paper do not involve very poor CME. According to Wang 2014,there is a discretion of manual operators during the detection of very poor cmes hence they are removed from the catalog in order to prevent any bias that may occur during the analysis. Yashiro et al. [2008 and Gopalswamy et al. [2010] reported inconsistency in the detection of CME with angular width ˂30o.Also the CME with angular width˃ 180o suffer from projection effect. Thus a lower threshold of 30o and an upper threshold of 180o is applied to the angular width of CMEs for the analysis.(6)

    Screenshot (309)_2.png
      Fig 2-Catalog made for the analysis from CDAW catalog

      The CME sources (AR, PE, AP), their spatial co-ordinates and rank of CME are noted in the catalog as shown in fig2.

      The three coronagraphs C1, C2, C3 of LASCO onboard SOHO provides the white light images of CME. But LASCO C1 was disabled in 1998 and hence only LASCO C2 & C3 provides the necessary data.In this paper LASCO C2 observations are made.

      LASCO occults the solar disc and creates a scenario of an artificial eclipse covering the brightest region of sun and helps to observe the eruptions from the solar surface. LASCO provides the white light images of CME.Wang 2002, proposed an identification method for the detection of CME source location on the basis of time and position angle (or propogational direction) as follows, EIT images are backplotted on LASCO, the LASCO movies are run for the specified time and position angle as listed in the SOHO LASCO CDAW catalog and the source regions of CME are roughly located on the EIT images. Accordingly the data was acquired within a temporal range of ±5 min and a spatial range of ±20o.

      The CME in solar cycle 24 were observed by back projecting LASCO images on Atmospheric Imaging Assembly, AIA 171 Å & AIA 304 Å onboard Solar Dynamic Observatory SDO. For the observations of CME sources in solar cycle 23, EIT 195 Å images were used since SDO was launched in 2010. SDO has 4 times greater resolution than SOHO and higher cadence. The Cadence of AIA 171 Å &A1A 304 Å is 12 s while that of EIT 195 Å is 12 min.

      The CME due to active regions (AR) were identified in EIT 195 Å in solar cycle 23 & AIA 171 Å for solar cycle 24. While the CME due to eruptive prominences (PE & AP) were observed in higher wavelength AIA 304 Å, Fig 3.

      For CME whose source regions were on the backside of the solar disc, observations were made using twin satellites Solar TErrestrial RElations Observatory (2006),STEREO A & STEREO B. Sun Earth Connection Coronal and Heliospheric Investigation (SECCHI) is an instrumental package of STEREO. SECCHI EUVI( Extreme Ultra Violet Imager) is one of the forur instrument of SECCHI. SECCHI EUVI 171 Å & SECCHI EUVI 304 Å for AP & PE identification.

      The JHelioviewer is a visualisation software which is used to visualise the sun for any time period from September 1991.They can perform the basic image processing in real time and track features on the sun. All the data are collected using JHelioviewer.

        Fig 3a-AR observed in AIA 171 Å
        Screenshot (320).png
          Fig 3b-PE observed in AIA 304 Å
          Screenshot (325).png
            Fig 3c-Filament which is embedded in active region
            Screenshot (324).png
              Fig 3d-AP resulting from the eruption of the filament, observed with AIA 171 Å & 304Å

              Also the type of CME is identified as shown in fig 4.The identification of CME is very essential for studying the kinematics of CME.A1 is termed to be a very good CME as it has a bright detectable leading edge which gives prominent information regarding the parameters of CME such as height etc. The A2 & B1 CMEs are good ones.B1 CME is a poor one as the leading edge gets distorted after travelling a certain distance. C class CME is a poorest one which is dispersed and hence is very week to study the kinematics. Quality of CME is seen in LASCO C2, it is better observed using the running difference images.

                Fig 4a- A1
                  Fig 4b- A2
                    Fig 4c- B1
                      Fig 4d- B2
                        Fig 4e- C

                        Challenges faced

                        1. The LASCO, EIT & AIA data was not available for few events which made it difficult to identify the exact source regions on the solar disc.

                        2. For solar cycle 23 high cadence images were only seen in EIT 195 Å, but very poor cadence was seen in EIT 171 Å & EIT 304 Å(6 hours cadence), hence detection of eruption processes were difficult.

                        3. Also SDO as well as STEREO images were not available for solar cycle 23. The lack of STEREO data caused difficulties while identifying the source regions which had their origin in the backside of the solar disc.

                        4. For certain CME events more than one activity or eruption was observed in on the solar disc as seen from AIA or EIT images, which caused difficulty in identifying the exact source location of the CME.


                        Fractional Contribution of Source region of CME

                        The fractional contributions of different source regions of CME for the data collected in solar cycle 23& 24 is shown in table 1,2&3. The AR source has greater contribution for the CME in both solar cycles 23 & 24 as it can be seen from table 1,2&3, the contribution seems to be correct as the observations are done in the rising phase of both the solar cycles when the activity of the sun is seen to increase from minimum to a maximum. Thus increasing activity leads to increase in AR. AP contributes least to the CME, as eruptions from AP are highly energetic and hence rarely observed. The contribution of PE is fairly seen but less than AR which can clearly be understood i.e.as the activity increases the quiescent area on the sun decreases. In fig 1, the activity of solar maxima in SC 23 is greater than that in SC 24, accordingly from the table we can notice that the AR contribution is larger in solar cycle 23 than that in solar cycle 24. The analysis is made for the entire set of observations made for solar cycle 23 & 24. A conclusion can be drawn that majority of CME have an onset from the active regions i.e. AR.

                        Solar cycle 24 (sept-dec 2010:jan-june2011)
                        Sl no  Source regionNo of events  Fractional contribution
                         1AR 217 52.80%
                         2PE  160 38.93%
                         3 AP 34 8.27%
                        Total number of events observed-411
                        Solar cycle 23 (jan-may 1998: march-june 1999)
                        Sl no  Source region No of events Fractional contribution
                         1 AR 249 72.17%
                         2 PE 72 20.87%
                         3 AP 24 6.95%
                        Total number of events observed-345
                        Solar cycle 23 & 24
                        Sl no  Source region No of events Fractional contribution
                         1 AR 466 61.64%
                         2 PE 232 30.69%
                         3 AP 58 7.67%
                        Total number of events observed-756

                        The magnetic flux observed in the photosphere is complex and reconfiguring continuously leading to the ejection of solar mass as solar flares, CME etc., this might be the explanation for the abundance of contribution of AR.

                        Latitudinal and longitudinal distributions of source regions of CME

                          Fig 5a

                          The Latitudinal distribution of AR for solar cycle 23 & 24 with bin size 3 o are shown in figure 5a.The distribution has a clear bimodal appearance and is symmetric about the equator. The plot obtained is bimodal within ±50o latitude from the equator, depicting that the sources AR are within the equatorial belt of on the photosphere.

                            Fig 5b

                            The longitudinal distribution of AR for solar cycle 23 & 24 with bin size 8 o is shown in figure 5b. There is no uniform distribution of longitude for AR but the source distribution increases with increasing absolute longitude. There is no east-west symmetry. The non-uniform distribution of the longitude suggests that the CMEs originating from solar limb could be observed more easily than those near the disk centre (4). The peaking of the data plot is seen at about 90o. In solar cycle 23 only the sources on the front side of the solar disc is observed as the STEREO was launched in 2006. In solar cycle 24 the longitudinal distribution beyond 90o is due to STEREO data.

                            FR 3.PNG
                              Fig 5c

                              The prominences are spread to all latitudes toward the solar maximum while during solar minimum PEs are confined to active region belt (2). The latitudinal distribution of PE for solar cycle 23 & 24 is shown in with bin size 8 o fig 5c The contribution of PE is less for both the solar cycles as it can be seen from table 1,2&3.There is no any symmetric bimodal distribution seen for solar cycle 23 from fig 4c but for solar cycle 24 a poor bimodal distribution is seen. CME due to PE can originate from almost all latitude(8). The non-uniformity in PE latitude distribution may be seen because the Prominences after rising to a certain height may be accelerated with other associated events like Solar flares creating an extra impulse for the prominence.

                              fr 4.PNG
                                Fig 5d

                                The longitudinal distribution of PE with bin size 10 o is shown in fig 4d, There is no uniformity seen in the distribution of the longitude but the peaking is observed at 90 o i.e. majority of the prominences erupted from higher longitudes away from the equator.

                                Latitude vs PA plots

                                Latitude vs PA plot gives the information regarding any deflections in the CME ejected, i.e. deflections of CME seen in LASCO FOV from their source regions observed in AIA OR EIT. From the plot shown in fig 6 no deflections are seen in the CME as a clear co relation between latitude and PA can be seen from the plots. This is actually due to lack of sufficient events identified.

                                FR 5.PNG
                                  Fig 6a
                                  FR 8.PNG
                                    Fig 6b

                                    FR 6.PNG
                                      Fig 6c
                                      FR 7.PNG
                                        Fig 6d

                                        In fig 6a & 6b there is clustering of events seen in the mid-latitudes in the range of ±50o which is similar to the latitudinal distribution seen from fig 6a, this results to the fact that no deflections is seen in AR for the events observed.

                                        The prominence originates from the active region belt and moves toward the equator, while the overlying field lines, which become the CME frontal structure, span the equator and hence appear at a very small PA.(8). The difference in distribution observed in PA vs latitude plots for AR and PE indicates the difference in spatial occurrences of them on the solar disc.

                                        CME width distribution

                                        In the CDAW catalog, the width of a CME is defined as the maximum angle subtended by a CME on the center of the Sun when the CME enters the C3 field of view (FOV) where the width appears to approach a constant value [Gopalswamy, 2004](6). For CMEs away from the limb, the measured width is likely to be an overestimate but it should be true width for CMEs originating from close to the limb. Here we are unaware of the fact that whether the width of CMEs for the data selected follows power-log distribution or log-normal distribution.

                                          Fig 7a

                                          The width distribution of AR & PE for solar cycles 23 & 24 are shown in fig 7. In solar cycle 24, the CME is distributed dominantly in the range 35 o to 65 o, whereas predominant distribution of width in solar cycle 23 is seen in a slight wider range 35 o to 80 o.

                                          FR 10,.PNG
                                            Fig 7b

                                            For PE, in both cycle 23 & 24, the width distribution is dominant at 40 o.

                                            Distribution of Position Angle (PA)

                                            The Position Angle of CME is the angle, measured counter-clockwise from the solar north pole, marked by the midpoint of the measured angular extent of the CME front (10). From fig 8 it can clearly be seen that the PA is predominantly distributed at 90 o & 270 o i.e. in the equatorial latitude. This implies that majority of the CME observed are found to eject at PA 90 o & 270 o which are from AR.

                                            FR 11.PNG
                                              Fig 8a
                                              FR 12.PNG
                                                Fig 8b

                                                The distribution of Position angle of AR & PE for solar cycle 23 & 24 is shown in fig 8.

                                                Fast and slow CMEs

                                                CMEs can be classified as fast or slow CME based on their speed relative to the speed of solar wind. According to Schwenn(9), the slow solar wind has the flux speed less than 400km/s(250-400km/s) while fast solar winds have the flux speed less greater than 400km/s (400-800 km/s). Thus for the statistical studies, the average speed of solar wind is considered as 400km/s. We classify CME with speed less than 300 km/s as slow CME and those with speed greater than 500km/s are considered as fast CMEs. The CMEs with intermediate speed between 300km/s to 500km/s are not considered in this paper (as the intermediate CME cannot be classified as fast or slow CMEs)(6). From the table 4,5&6, it can be seen that for the data collected in solar cycle 23, the contribution of fast CME is seen to be slightly more(53.46%) than the slow CMEs(46.51%). Whereas in solar cycle 24 majority of slow CMEs are observed (70.19%) and very few slow CMEs(29.81%) are observed. So no exact conclusion can be drawn regarding the contribution of fast and slow CMEs, but the overall contribution of slow CME is seen to be slightly high (59.58%) than that of fast CMEs (40.42%).

                                                Solar cycle 23
                                                 Sl no CME type No of events Fractional contribution
                                                 1 Slow CME 100 46.51%
                                                 2 Fast CME 115 53.49%
                                                Total No of fast & slow CMEs observed-215
                                                Solar cycle 24
                                                 Sl no CME type No of events Fractional contribution
                                                1  Slow CME 186 70.19%
                                                 2 Fast CME 79 29.81%
                                                Total no of fast & slow CME observed-265
                                                Solar cycle 23 & 24
                                                 Sl  no CME type No of events Fractional contribution
                                                 1 Slow CME 286 59.58%
                                                 2 Fast Cme 194 40.42%
                                                Total no of fast & slow CME observed- 480

                                                Table 4,5 & 6 gives the fractional contribution of Fast and slow CMEs in solar cycle 23 & 24 for the data selected.

                                                FR 13.PNG
                                                  Fig 9a

                                                  The distribution of slow CMEs for solar cycle 23 & 24 can be seen in fig 9a.In solar cycle 24, very few CMEs with speed˂100km/s are seen. The distribution increases gradually for speed ˃100km/s and decreases. Maximum CME are observed with the speed ̴210 km/s. For solar cycle 23, there is a poor contribution of CMEs with speed ˂140km/s. Similar to solar cycle 24 an increase is seen in distribution beyond 140km/s and a decrease. Maximum CME are observed with speed ̴270km/s.

                                                  FR 14,.PNG
                                                    Fig 9b

                                                    In the distribution of fast CMEs, fig 9b, there is an exponential decrease seen in the distribution. Maximum CMEs are seen to have the speed ̴500km/s in solar cycle 24 and speed ̴600km/s in solar cycle 23. Very fast CMEs with speed˃100km/s are hardly seen.

                                                     CME width vs PA distribution

                                                      Fig 10a
                                                        Fig 10 b
                                                        FR 18.PNG
                                                          Fig 10 c
                                                          FR 17.PNG
                                                            Fig 10 d

                                                            In fig 10a & 10c, clustering of the active regions can be observed in the mid-latitude regions i.e. in the equatorial active region belt, while no such clustering is observed in the PE.

                                                            Similar analysis of the properties of CME, which are mentioned in this paper, can be carried out for CME from A, but the AP data collected is very less as it can be clearly witnessed from tables 1,2&3. Thus due to lack of data, analysis of AP is ignored.


                                                            Coronal mass ejection is a process by which sun losses it mass in the form of plasma associated with magnetic field, from the corona. The CMEs have their origin on the photosphere, either active regions or filaments. The identification of the source region CME and studying the associated properties gives good information regarding the activity of the Sun. In this paper the source regions of 756 CMEs were identified in the rising phases of solar cycle 23 & solar cycle 24 and their various associated properties were analysed as follows

                                                            1. The latitudinal and longitudinal distribution reveal about the spatial location of the sources of CME. It is found that AR are seen in the mid latitude range depicted by the bimodal distribution ranging in ±50 o, while that of PE has no clear bimodal distribution, indicating that PE for the data collected is spread in the wide latitudinal range. From longitudinal distribution of AR as well as PE it is clear that Majority of the CME have ejected from the longitude 90o

                                                            2. From the plot of Latitude vs PA, we found no as such deflection in CME is observed for the collected data.

                                                            3. We found maximum distribution of CME width in the range 30 o -50 o

                                                            4. Majority of CMEs were found to have PA at 90 o and 270 o i.e. in the equatorial belt.

                                                            5. We have analysed the contribution of fast and slow CMEs and also their distribution.

                                                            6. We have found the clustering of AR in the mid-latitude from the distribution of CME width vs PA, while PE are found to be scattered.

                                                            FUTURE WORK

                                                            Data can be collected for the entire solar cycle 23 & 24 at different phases(minima, maxima, raising & falling phases) & the comparative studies can be done for the various statistics obtained in both the solar cycle and the prediction of the activity of the solar cycle 25 can be made, especially to understand the solar minima of solar cycle 25 which we are presently undergoing. Along with the activity, the other characteristics like the rate of increase and decrease in the activity in the rising and falling phases of solar cycle 25 can be looked at. The width and PA distributions also provide fruitful information in the prediction. Thus the study can be continued for forecasting the upcoming solar cycle.


                                                            1. Timothy Howard, Coronal Mass Ejection-An Introduction

                                                            2. N Gopalaswamy(2004), A Global picture of CMEs in inner Heliosphere

                                                            3. Prasad Subramanian, K. P. Dere -Source Regions Of Coronal Mass Ejections

                                                            4. Yuming Wang, Caixia Chen, Bin Gui, Chenglong Shen, Pinzhong Ye, and S. Wang-Statistical study of coronal mass ejection source locations: Understanding CMEs viewed in coronagraphs

                                                            5. N. Gopalswamy Æ S. Yashiro Æ G. Michalek Æ G. Stenborg Æ A. Vourlidas Æ S. Freeland Æ R. Howard-The SOHO/LASCO CME Catalog

                                                            6. V. Pant, S. Majumdar, A. Chauhan, D. Banerjee, N. Gopalswamy-Width distribution of fast and slow CMEs in solar cycle 23 and 24: Power-laws or Log-normals, that is the question!

                                                             7. Bin Gui · Chenglong Shen · Yuming Wang · Pinzhong Ye · Jiajia Liu · Shui Wang · Xuepu Zhao-Quantitative Analysis of CME Deflections in the Corona

                                                             8. N. Gopalswamy, M. Shimojo, W. Lu, S. Yashiro, K. Shibasaki, and R. A. Howard- Prominence eruptions and coronal mass ejection: a statistical study using microwave observations.

                                                             9. Rainer Schwenn-Space Weather: The Solar Perspective

                                                            10. Alejandro Lara, The Source Region of Coronal Mass Ejections, The Astrophysical Journal · December 2008


                                                            I Kavya Bindu M, express by sincere gratitude to my mentor Prof Dipankar Banerjee for all the knowledge, support and guidance he has given me. I consider it as privilege on my part for having given an opportunity to perform a project under his guidance.

                                                            My sincere thanks also go to the co mentors, Mr. Satabdwa Majumdar and Mr. Ritesh Patel, for helping me to carry the project and guiding me in every step.

                                                            I thank the Indian Academy of Sciences, Bangalore for providing me with an opportunity to intern at IIA through Summer Research Fellowship Program 2019. I also thank the Indian Institute of Astrophysics, Bangalore for providing me with all the resources required for the project.

                                                            Written, reviewed, revised, proofed and published with