Comparison Between Radio Loud and Radio Quiet Fast CMEs: A Reason for Radio Quietness

It is well known that fast CMEs are mostly associated with magnetohydrodynamic (MHD) shocks in the solar corona, forming type-II radio bursts. However, the absence of type-II radio bursts is not uncommon. Herein, we aim to analyze the differences between the radio loud (RL) and radio quiet (RQ) fast Coronal Mass Ejections (CMEs) (speed ≥ 900 km s−1) during Solar Cycle 24 (2008 – 2021). From the 309 fast CMEs, we could identify 143 events with a known source origin on the visible disk (Earth view). We identified the associated flares/CMEs for 143 events using running-difference images from (i) Solar Dynamic Observatory/Atmospheric Imaging Assembly (SDO/AIA) and (ii) Large Angle Spectrometric Coronagraph (LASCO) observations. Among these 143 events, RQ and RL groups have 70 and 73 events, respectively. CALLISTO and Wind/WAVES observations are used to identify these RL and RQ sets. We analyzed the possibilities of streamer-CME and CME-CME interaction. In this study, we report the important differences between RL and RQ CMEs and the underlying reasons for the radio quietness of fast CMEs. In the LASCO field of view, the majority of RL CMEs (almost 90%) interacted with streamers and/or pre-CMEs, whereas only 25% of RQ CMEs did the same, and there was no pre-CME interaction. The observational evidence led to the conclusion that substantial density perturbation/interaction increases the probability of production of type-II radio emissions by the shock of RL CMEs.

Type-II radio bursts are slow-drifting bursts, which are signatures of outward motion of MHD shocks in the solar corona (Payne-Scott, Yabsley, and Bolton, 1947;Wild and Mc-Cready, 1950;Wild, Murray, and Rowe, 1953;Uchida, 1960;Wagner and MacQueen, 1983;Nelson and Melrose, 1985;Gosling, 1993;Dryer, 1996;Magara et al., 2000;Cliver et al., 2004;Shanmugaraju et al., 2014;Gopalswamy, 2016;Syed Ibrahim et al., 2021). Coronal and Interplanetary shocks are produced by fast CMEs whose speed exceeds the local magnetosonic speed. These shocks are accelerating electrons in the ambient medium. Type-II burst reveal the formation height of CME associated shock and its acceleration (Knock et al., 2001;Mann and Klassen, 2005;Cho et al., 2008;Gopalswamy et al., 2009Gopalswamy et al., , 2013aRamesh et al., 2010;Cho et al., 2011;Ramesh et al., 2012;Shen et al., 2013;Umuhire et al., 2021). Metric and decameter hectometer (DH) type-II radio bursts associated with CMEs are highly energetic events compared with other CMEs not associated with type-II radio bursts. Metric and DH distance ranges are 1 -2.5 R and 2 -10 , respectively (Gopalswamy et al., 2005). CMEs associated with type-II radio burst are called Radio Loud (RL) CMEs, and those which are not associated are called Radio Quiet (RQ) CMEs (Sheeley et al., 1984;Gopalswamy, 2004;Michalek, Gopalswamy, and Xie, 2007;Gopalswamy et al., 2008). Michalek, Gopalswamy, and Xie (2007) discovered the angular size distinction between RL and RQ CMEs. They assembled these two samples of occurrences for this purpose using data from Wind/WAVES and SOHO/LASCO observations made between 1996 and 2005. They demonstrated that the width of the RL CMEs is about two times that of the RQ CMEs (considering expanding parts of CMEs). The RQ CMEs also have a bright broad expanding region with a significant extended diffusive structure. According to Gopalswamy et al. (2008), from 1996 to 2005, almost 40% of the fast wide (FW) CMEs are associated with RQ cases. The average speed of RQ CMEs was 1117 km s −1 , whereas the RL's average speed was 1438 km s −1 . The proportion of halo CMEs (apparent width = 360 • ) was highest for the RL CMEs (60%), lowest for the RQ CMEs (16%), and intermediate for the FW CMEs (42%). The front-sided CMEs originated close to the limb, while the back-sided ones comprised the majority of RQ CMEs (around 55%). Only around 25% of the RL CMEs had backside origins; the majority were on the disk.
The properties of CMEs connected to long-wavelength type-II radio bursts in the near-Sun interplanetary medium were studied by Gopalswamy et al. (2001). They observed a deceleration for fast CMEs (speed ≥ 900 km s −1 ). They suggested that the coronal drag could be responsible for the deceleration, based on the result that the deceleration has a quadratic dependence on the CME speed. About 60% of the fast CMEs were not associated with DH type-II bursts, proposing that some additional condition needs to be satisfied to be RL. The average width (66 • ) of the RQ fast CMEs is much smaller than that (102 • ) of the RL CMEs, suggesting that the CME width plays an important role. The special characteristics of the RL CMEs reveal that the detection of DH radio bursts may provide a useful tool in identifying the population of geo-effective CMEs.
Additionally, Mäkelä et al. (2011) looked into the variability of energetic storm particle (ESP) events related to shocks caused by CMEs. The interplanetary shocks were detected between 1996 and 2006. They examined the CME's characteristics close to the Sun. The CMEs with an ESP-producing shock are faster (V = 1088 km s −1 ) than those (V = 771 km s −1 ) shocks without an ESP event, and they have a higher percentage of halo CMEs (67% versus 38%). Further, Gopalswamy et al. (2012) examined a group of 21 type-II bursts with CME sources located within 30 • of the disk's center that were detected by the Wind/WAVES experiment at decameter-hectometric (DH) wavelengths but lacked a shock at 1 AU. They discovered that these CMEs' typical near-Sun speeds are only 644 km s −1 , slightly faster than the average speed of CMEs connected to RQ shocks. However, the fraction of halo CMEs is just 30%, as opposed to 54% for RQ shocks and 91% for all RL shocks.
Recently, Suresh and Shanmugaraju (2015) analyzed the differences between RL and RQ associated CMEs from 1996 to 2012. They investigated 552 flare-CME pairs (those associated with RL and RQ CMEs) from the set of 891 fast CMEs (speed ≥ 900 km s −1 ). They tried to understand the reason for the radio quietness of some fast CMEs. They found that 38% fast CMEs are associated with type-II radio bursts, and 62% fast CMEs are not. The result showed that RQ events have a poor association with halo CMEs (∼10%), but RL events are highly correlated with halo CMEs (∼66%). The mean acceleration/deceleration and initial acceleration values of RL CMEs are slightly higher than RQ CMEs. They found that RQ CMEs are less energetic than RL CMEs.
In the present work, we selected 309 CMEs with speed ≥ 900 km s −1 . Gopalswamy et al. (2008) stated the reason for this selection criteria. They explained that a significant fraction of high-speed CMEs were not associated with type-II radio bursts because some CMEs could not drive the shock, so type-II observations were impossible. Otherwise, these radio waves do not propagate in the Sun-Earth line. According to Michalek, Gopalswamy, and Xie (2007), the RQ events are not very wide due to poor correlation with halo CMEs. The percentage of halo CMEs serves as a measure of kinetic energy and speed. It brings up the original argument for why fast and wide CMEs are RQ, so we initiated this study. In this analysis, we tried to provide a suitable reason for radio quietness. Here, we analyzed the differences, in a somewhat different perspective from the earlier studies, between RL and RQ fast CMEs (speed ≥ 900 km s −1 ) during the complete Solar Cycle 24 (December 2008-August 2021). Data analysis is given in Section 2, and results are presented in Section 3. Finally, the summary and conclusions are given in Section 4.

Data Analysis
In the present study, we checked the CME-ejection location and the corresponding flare eruption region using (i) Solar Dynamic Observatory/Atmospheric Imaging Assembly (SDO/AIA; Lemen et al., 2012) and (ii) Large Angle Spectrometric Coronagraph (LASCO; Brueckner et al., 1995) running-difference images to identify the CME-flare pair. We focused on RL and RQ fast CMEs (speed ≥ 900 km s −1 ) during the Solar Cycle 24 (December 2008-August 2021).
There were 309 CMEs with velocities ≥ 900 km s −1 between 2008 and 2021. After excluding the backside events (162), we obtained 147 flare-CME pairs. After that, we checked the following instruments (i) Radio and Plasma Waves experiment on board the Wind spacecraft (Wind/WAVES, Bougeret et al., 1995) and (ii) extended Compound Astronomical Low frequency Low cost Instrument for Spectroscopy and Transportable Observatory (CAL-LISTO, Benz, Monstein, and Meyer, 2005) to exclude any data gap for every RQ events. We excluded 4 CMEs from the RQ list because of lack of observation in Wind/WAVES. Finally, we obtained 143 flare-CME pairs. In this study, the main difference is that we did not use any other selection criteria but only confirmed each flare-CME association using SDO/AIA data and LASCO CME running-difference images. We also tried to cover the whole Solar Cycle 24 period. Flare soft X-ray and CME details are obtained from GOES flare catalog 1 and LASCO CME catalog, 2 respectively. In addition, RL and RQ associated events are divided into two sets as accelerated and decelerated CME events.
We carefully checked the type-II observation data in the online Wind/WAVES catalog (developed by Gopalswamy, Mäkelä, and Yashiro, 2019 3 ) to confirm the association between the CME and type-II radio bursts. There is only one new event (originated on 7 December 2021) in the present study apart from the Wind/WAVES type-II list. Metric and DH type-II radio burst observation details are obtained from National Oceanic and Atmospheric Administration (NOAA) 4 and Wind/WAVES observations. Wind/WAVES detected a number of type-II radio bursts, which are described in detail in the online catalog. The catalog has been improved by compiling the associated flares, CMEs, and solar energetic particle (SEP) occurrences along with their fundamental characteristics. Among the selected 143 events, for the present study, we got 73 RL CMEs and 70 RQ CMEs (see Table 1). First, the properties of CMEs and flares of the two groups (RL and RQ) are analyzed statistically and then compared. Second, their associated activities, such as metric type-IIs and CME-streamer or CME-CME interaction, are studied.
Examples are shown in Figures 1 and 2. LASCO C2 and C3 observations are shown in the top panel, and the bottom panel shows the dynamic radio spectrum from Wind/WAVES. Figure 1 shows RL CME observed on 7 March 2012 at 00:24 UT. It was observed in the height range of 2.36 to 28.69R . Corresponding observations of Wind/WAVES type-II/III are indicated with arrows for the RL CME. Similarly, we have shown the LASCO C2, C3, and Wind/WAVES observation for the RQ CME in Figure 2. For this CME, there is no type-II radio observation in Wind/WAVES radio spectrum during the time of LASCO CME observation, but the instrument observed the CME associated type-III bursts. For the lack of the type-II radio burst observation, this CME was identified as RQ CME.

Flare and CME Properties
The heliographic distribution of RL (solid black circle) and RQ (solid red circle) associated events is given in Figure 3 using flare locations. We compared the previous result by Suresh Bottom panel: Wind/WAVES dynamic radio spectrum observations corresponding to the CME. Type-II and type-III radio burst signatures are indicated by the arrows. and Shanmugaraju (2015) and found notable differences in the present study: (i) large number of RQ events (42/70) occurred in the limb side (longitude ≥ 60 • ) of the Sun and (ii) small number of RL (29/73) events occurred in the limb side of the Sun. Suresh and Shanmugaraju (2015) noted a sharp cut-off in the latitudinal distribution around ±30 • . However, 5 events (4 RL and 1 RQ) occurred in the current analysis over 30 • latitudinal regions but below 45 • . This tiny variance is irrelevant and has no impact on the study's findings or conclusions. We also found that most of the RL events occurred in the northern hemisphere in Solar Cycle 24.
Among the 73 RL events, more cases are also associated with M-and X-class flares (79%) than the RQ events (39%). However, nearly 15% of the RQ CMEs are only associated with less energetic B-class flares. Our studies confirmed that the accelerated RQ CMEs originated from M-and C-class flares, but there is weak association with decelerated RQ CMEs and X-class flares (see Table 2). The mean values of RL CME parameters (like angular width, linear speed, initial speed, and the number of halo CME occurrences) are higher than those of RQ CME parameters. The statistical values of the CME parameters are comparable to found in previous studies (see, e.g., Vršnak, 2008;Ramesh et al., 2010 andPrakash et al., 2012). Recently, Jang et al. (2016) studied and compared two-dimensional (2D) and three-dimensional (3D) CME speeds determined using observations by SOHO and Solar TErrestrial RElations Observatory (STEREO). They obtained 306 LASCO front side CMEs from 2009 to 2013. They found that 2D speeds are underestimated to the 3D speed by about 20%. In the present work, we used their estimated CMEs' 3D speed values to find the aver- Figure 2 Top panel: Radio Quiet CME observed in LASCO C2 and C3 field of view on 2011 August 9. Bottom panel: Wind/WAVES dynamic radio spectrum observations corresponding to the CME. Type-III radio burst signatures are indicated by the arrows. Type-II radio burst signatures are not observed for quiet CMEs. age 3D speed for RQ and RL CMEs. We found that the RL CMEs' 3D speed (1717 km s −1 ) value is much higher than the RQ CMEs (1368 km s −1 ). From this section, we conclude that the reasons for radio quietness of the CMEs may be the following: (i) every RQ CME has less energetic flare association, and (ii) short rise and decay time of flares (∼60 minutes). Some more points have already been discussed earlier by Gopalswamy et al. (2008) in this connection.

Interaction of RQ/RL CMEs with Streamers and Pre-CMEs
Further, we analyzed CME-CME and CME-streamer interaction possibilities for both RQ and RL cases. In the RQ cases, the number of CME-streamer and pre-CME interactions is found to be considerably lower/smaller than in the RL CMEs. Nearly 52 RQ CMEs have no streamers in the LASCO field of view. Further, there are also no pre-CME interactions and CME-steamer interactions. The remaining 18 CMEs have streamer interaction. Among the   Gopalswamy et al. (2008). In case of RL CMEs, most of the events are associated with streamer and pre-CMEs: among the 73 RL CMEs, 22 events seem to have interacted with streamers, and 45 events seem to have interacted with the tail of pre-CMEs. Examples of streamer and pre-CME interactions are shown in Figures 4 and 5, respectively. It seems that shock formation probability increases because of the density perturbation. We note from Figure 4 that streamer signatures were visible in the LASCO observations before the CME eruption (indicated by the multiple black arrows). Also, RL CME interacted with the pre-CME eruption, as shown in Figure 5. Here post-CME has interacted with the tail of pre-CME. The first CME was ob-Figure 4 RL CME observed on 2013 September 29 in LASCO C2 field of view. This RL CME has interacted with streamer. The arrows indicate the position of CME and streamers. served on 22 May 2013 8:48 UT, then after 4.5 hours, we noticed the second CME erupted from the same quadrant of the Sun. The mean position angle is nearly the same for both CMEs. Hence in these cases, the streamer and pre-CME interactions increase the probability of type-II radio burst formation. Kong et al. (2015) analyzed two coronal type-II events that erupted from the same AR below a streamer. They confirmed the formation of type-II emission when the CME front crossed the streamer structure and ended around the time the CME/shock front passed by the white-light streamer. In the present study, we observed that 85% of interaction cases are associated with type-II bursts.
For further confirmation, we observed a link between the CME interaction time (in LASCO FOV) and the type-II formation time in the DH range. We did this for RL events by calculating the delay time between CME interaction and type-II formation (see Figure 6). The resulting delay time is extremely short, confirming the findings of a strong relationship between type-II formation and interaction. In the LASCO FOV, we measured the CME's first interaction height. Table 3 provides the relevant information. We can see the interaction height, error in height measurement, and delay time between the type-II and interaction start time in this table. We examined the following CME parameters of post-and pre-CMEs for CME-CME interaction using LASCO data: (i) mean position angle, (ii) angular width, and (iii) CME speed. Here, the leading edge of the post-CME mostly interacted with the tail of the pre-CME. The streamer deflection during the CME eruption was also examined for the CME-streamer interaction. We repeatedly measured the CME interaction height from the Sun's edge, and then we averaged all of the measurements for each particular event. Afterwards, we found the deviation of every single measurement from the average value. Finally, we gave the average deviation value as error in the interaction height in Table 3. Interaction of the events is confirmed by the help of central position angle and deflection of the events. Examples of CME interaction region, pre/post CME eruptions, and CME deflection are all displayed in Figure 7.  Heights computed using the DH type-II burst emission are compared to the interaction height in Figure 8. The solid circle shows the heights using the Leblanc one-fold density model between 2 and 7R (Leblanc, Dulk, and Bougeret, 1998;Leblanc et al., 2001). The empty circles indicate the height of the interacting CMEs with pre-CMEs or streamers. Also, the model derived by Gopalswamy et al. (2013a) is shown as a dashed line. We can see from Figure 8 that the heights estimated using density models and the CME interaction height have a strong correlation. The result presented above that the interaction increases the probability of a shock forming type-II radio burst in the vicinity of the Sun is concluded. The coronal electron density is a major factor in determining the type-II emission frequency, and any variations in coronal density along the shock path will have a better ambient for producing type-II radio emission. Since CME and shock interact with streamer, it is widely recognized that this region is an important source for type-II bursts (Sheeley, Hakala, and Wang, 2000;Cho et al., 2008Cho et al., , 2011Chen et al., 2010;Feng et al., 2011;Shanmugaraju et al., 2017).
In all the above analyses, the interaction was validated using LASCO white-light running-difference images. After the coronagraph occulter disk, we documented the com-  mencement time of the CME collision. Perhaps the CME interaction began at a lower height obscured by the occulter disk. Because of the time gap (12 minutes) between LASCO image sequences/observations, we may miss the exact interaction height.

Figure 7
RL CME was spotted on 2011 February 24 using LASCO C2 and C3 coronagraphs (a)-(d). Interaction height of the CMEs, pre/post CME eruptions and deflection of the CME are mentioned.

Figure 8
The height estimated from the starting frequency of DH type-II observations is compared to the height of the CME's interaction with any pre-CMEs or streamers. The Leblanc's one fold density model for heights between 2 and 7R is shown by the solid circle. The height of the interacting CMEs is indicated by the empty circles. 29 events out of the 73 RL CMEs were linked to a starting frequency of 16 MHz, whereas 25 events were linked to a starting frequency of 14 MHz. In these events, we took the average single height value for frequency 16/14 MHz for the same frequency start. These two values are indicated by red circle with star. The frequency-height relation from Gopalswamy et al. (2013a) is drawn as dashed-line.

Summary and Conclusion
Initially, we selected 309 events with CMEs speed ≥ 900 km s −1 . After that, we checked the corresponding flares and found 143 events using running-difference images. Among the 143 events, we identified that 70 and 73 are associated with RQ and RL cases, respectively. From the overall comparison between RL and RQ events, we found that RL CMEs are highly associated with M-and X-class flares (79%) compared to RQ CMEs (39%). Using the flare observation by GOES data, we also tried to find the reason for radio quietness of fast CMEs. In the RQ case, 14% were associated with B-class flares. However, all RL events were associated with C-, M-, and X-class flares. CME acceleration and deceleration were observed in large numbers in both RQ and RL events. These results reveal the known fact that the RL CMEs are more energetic and associated with stronger and longer duration flares than the RQ CMEs. Moreover, many of the RQ events have occurred at large angles from the Sun-Earth line. More importantly, we checked the streamer and pre-CME interaction for the RL and RQ cases. We found that most of the RL CMEs (nearly 90%) interacted with streamer or pre-CMEs in the LASCO field of view, but only 25% of RQ CMEs interacted with streamers, and there was no pre-CME interaction in the LASCO field of view. This observational evidence concludes that shock formation probability is increased for RL CMEs because of strong density perturbation/interaction.