By Paul Roggemans, Carl Johannink and Takashi Sekiguchi
Abstract: A case study has been made on the available data for the h Virginids (HVI#343) and a reliable long-term reference orbit has been calculated. The orbits obtained by different independent video camera networks during the enhanced activity in 2020 are in perfect agreement. The orbit is a typical Jupiter family comet orbit. The HVI-meteor shower is likely a component of a complex which include some other known meteor showers.
The different CAMS networks registered enhanced activity from the h Virginids (HVI#343) between April 23 and 28 when Dr. Jenniskens decided to issue a CBET (Jenniskens, 2020), to announce the exceptional activity of the shower. Also, the SonotaCo Network in Japan and the Global Meteor Network could immediately confirm the enhanced activity of this shower.
The enhanced activity level lasted several nights more. The median values for the orbital elements of 13 orbits as calculated by Jenniskens (2020) were:
- q = 0.739 AU
- a = 2.99 AU
- e = 0.755
- i = 0.71°
- ω = 67.9°
- Ω = 216.9°
- Π = 284.5°
Which corresponds to a typical orbit of a Jupiter family comet. Using these values as reference orbit the CAMS BeNeLux network alone had 51 meteors that could be matched with this reference orbit until April 27–28. The exceptional favorable weather circumstances during this period provided perfect conditions to collect orbits during most of the activity period.
2 What do we know about HVI#343?
The shower has the status as established and the IAU working list of meteor showers (Jopek and Jenniskens, 2011; Jopek and Kaňuchová, 2014, 2017) mentions three references as evidence for this meteor stream. The first refers to SonotaCo Network (2009), a publication based on the first two years of the Japanese network (Table 1). Hence the discovery of the h Virginids is on the account of the SonotaCo Network.
The second reference in the MDC list is based on single station video work (Molau and Rendtel, 2009). The radiant position differs 10°, the activity period is about a week earlier and the velocity estimate is well above the known value for the HVI shower. The high risk for chance line-up contamination with non shower meteors and the presence of other nearby radiants are possible explanations for these differences. As the data is not based on orbits we do not consider this entry as a relevant reference for this shower.
The only evidence with a reference orbit in the IAU shower list comes from Jenniskens et al. (2016). This reference orbit was obtained from 11 HVI#343 orbits, all recorded between 27 April 2012 and 2 May 2012. The orbital elements of this reference correspond indeed with the median values for the 11 orbits, but we obtain different values for the solar longitude and radiant with λʘ = 40.2°, αg = 206.4° and δg = –12.0°. Working with median values to obtain an average orbit is not ideal, it is not the best way to average angular values. In this particular case with 11 orbits, 7 are south of the ecliptic, 4 are north of it, which means the nodes and argument of perihelion switch with 180° whether the orbit is north or south of the ecliptic. Therefore we applied the method of Jopek et al. (2006) to compute the average orbit for these 11 orbits. This average orbit is listed in Table 1 to be compared to the original reference. Checking for HVI orbits in the 2010–2016 CAMS orbit dataset (Jenniskens et al., 2018) which was released begin 2020, we find as many as 77 HVI orbits listed. Strange enough, only 3 of the 11 orbits that were identified as HVI in the 2010–2013 CAMS orbit dataset are still identified as HVI orbits in the new dataset. No new reference orbits have been published for the CAMS data. We compute the mean orbit for the 77 HVI orbits listed in the CAMS dataset 2010–2016, using the method of Jopek et al. (2006). The resulting orbit is listed in Table 1. This is a slightly different orbit than the one published by Jenniskens et al. (2016) which is listed in the IAU MDC. This explains why 8 of the original 11 HVI orbits failed to be identified in the newer dataset because a different reference orbit was derived for the shower.
Table 1 – The reference orbits listed for the HVI#343 as listed in the IAU working list of meteor showers compared to the average orbit for two CAMS datasets as computed by the authors.
et al. 2016
|vg||18.7 km/s||17.2 km/s||17.2 km/s||18.2 km/s|
|a||–||2.28 AU||2.26 AU||2.81 AU|
|q||–||0.742 AU||0.750 AU||0.766 AU|
It is obvious that apart from the SonotaCo network, the discoverer of this shower, the other references were somehow problematic. Masahiro Koseki (2020) wrote earlier: “HVI1 is quite different from others and forms possibly a different shower with 021AVB4 and 5, 136SLE2”. It is not excluded that the HVI#343 activity has been detected earlier in other surveys and got listed under a different shower identification.
2 How did HVI perform previous years?
We have 1101924 orbits public available, 630341 combined for EDMOND and SonotaCo (2007–2019), 471583 for CAMS (2010–2016). From the 2020 orbit data we know that h Virginids must be present within the interval 30° < λʘ < 46°, we limit our scope to this time interval which still contains 31443 orbits. The geocentric velocity range and the radiant position can help to reduce the size of the dataset further. Based on past HVI orbit data we estimate the useful velocity range to search HVI-orbits as 12 km/s < vg < 23 km/s. To limit the radiant area, we select a range in the Sun-centered ecliptic coordinates with 156° < λ – λʘ < 176° and –7° < β < +5°. This way the influence of the radiant drift due to the rotation of the Earth around the Sun is eliminated. This selection results in a workable dataset with 627 orbits.
We start a search on these 627 orbits to locate orbits that form a concentration using an iterative procedure (Roggemans et al., 2019). Using simple average values or median values for angular values is not recommended, moreover the HVI meteor stream has orbits north and south of the ecliptic what means that the nodes switch and the argument of perihelion differs 180°. We use the method of Jopek et al. (2006) to obtain average orbits for each collection of similar orbits.
To assess the degree of similarity between any orbit and the reference we use the so called discrimination criteria or D-criteria. To reduce the amount of contamination of the sample by sporadic false positives we apply the D-criteria of Southworth and Hawkins (1963), Drummond (1981) and Jopek (1993) combined. Each of these criteria has some slightly different eliminations. To distinguish between the weak similarity of the very dispersed part of the shower and the very strong similarity within the core of the shower, we define five different classes with specific threshold levels of similarity:
- Low: DSH < 0.25 & DD < 0.105 & DH < 0.25;
- Medium low: DSH < 0.2 & DD < 0.08 & DH < 0.2;
- Medium high: DSH < 0.15 & DD < 0.06 & DH < 0.15;
- High: DSH < 0.1 & DD < 0.04 & DH < 0.1.
- Very high: DSH < 0.05 & DD < 0.02 & DH < 0.05.
These classes are defined arbitrary and help us to evaluate the compactness of the collection of orbits. Old and very diffuse showers should contain mainly orbits with low threshold D-criteria while young or very compact showers should have a distinct core of very high threshold orbits. Many showers have a profile in between these two extremes.
We started with the average orbit as a reference orbit for all 627 orbits in our dataset. Using the low threshold class required four iterations to reach a reference orbit for a concentration of 528 orbits within our 627 available orbits. Within this selection we had 184 orbits that fit the high threshold class, but only 28 orbits within the very high threshold class. After another six iterations we got a reference orbit for the high threshold class with 188 orbits. This selection surprisingly included a core with 70 very high threshold orbits. We decide to use this selection to iterate the final reference orbits. After another six iterations we arrived at a final reference orbit valid for the core of the HVI meteor stream.
Working with the discrimination criteria requires caution. The results indicate only a degree of similarity between the orbits. D-criteria provide no proof for any physical relationship between the meteoroids. D-criteria can be very misleading, especially if applied on short period orbits with small eccentricity. In this specific case our iterations end with a very distinct reference orbit for the very similar orbits. Table 2 lists all the parameters obtained for the different threshold classes, using the final very high threshold orbit as reference.
Table 2 – The mean orbits for the HVI#343 meteor stream for each threshold class of similarity with the very high threshold average orbit as reference orbit.
|Medium High||High||Very High|
|vg||18.1 km/s||18.1 km/s||18.1 km/s||18.1 km/s||18.0 km/s|
|Hb||92.6 km||93.4 km||93.9 km||95.0 km||95.3 km|
|He||82.1 km||82.4 km||82.7 km||83.5 km||83.8 km|
|a||2.57 AU||2.67 AU||2.75 AU||2.83 AU||2.91 AU|
|q||0.751 AU||0.757 AU||0.759 AU||0.763 AU||0.770 AU|
In the IAU MDC working list we find a mixture of orbits for many showers without any indication for the degree of similarity used to search for the orbits. Many authors defined shower associations using a single D-criterion like DSH (Southworth and Hawkins, 1963) with a far too tolerant limit of 0.25 which produces plenty of false positives for certain types of orbits. Such meteor stream searches of the past resulted in large numbers of phantom meteor showers, defined by pure chance fitting sporadics.
Another information missing in most meteor stream searches is the time range in which the shower orbits were detected. In Table 2 we indicate the begin and end of the interval in λʘ . 30° < λʘ < 46° is the time range we selected our orbits to search for HVI orbits. These are present from the beginning till the end, hence we took our interval rather short and the actual activity period of the h Virginids certainly extends before and after the interval we selected. However, it is not the purpose to identify the outliers of this stream. Meteor showers get dispersed and meteoroids can get separated from the main stream until the point that D-criteria or other tools of identification fail to determine their shower association.
Any indication for periodicity?
HVI orbits were detected in each year for which orbit data during the activity period was available. The largest number of HVI orbits was recorded in 2015 with 109 orbits, but the total number of video orbits available during the HVI activity period was also very high that year. If we look at the proportion HVI orbits as percentage of the total number of orbits available we see that the HVI shower appeared stronger in 2008 and 2019 than in other years. 2008 was the year the shower was first detected by SonotaCo Network, note the strong presence of high and very high threshold class orbits that year. For 2019 we have only SonotaCo Network orbits available and also this year had a significant proportion of HVI orbits. Good coverage of the activity period in the period 2011–2016 excludes that any enhanced activity was missed. Weather may have affected the years 2010, 2017 and 2018 when too few orbits were registered around the maximum of the shower. The strong presence in 2008 (Figure 2) may indicate a previous HVI outburst, the year the shower caught the attention when it was first discovered.
The activity profile
When we count the total number of orbits and the number of HVI orbits in the same time bin, we can get an idea of the activity profile. We use a sliding mean counting the numbers of orbits per degree in solar longitude, moving forward in time in steps of 0.25° in solar longitude. The percentage of HVI orbits relative to the total number of orbits is a good indication for the activity level of the HVI shower. To eliminate the influence of other major sources on the total activity, we remove the orbits identified as Lyrids or as eta Aquariids.
First thing we notice is that HVI orbits are distinct present from the beginning until the end of the interval we searched for these orbits. This means the activity period starts earlier and ends later. There is no indication for any sharp peak in Figure 3, but two peaks are clearly visible, a first at λʘ = 36.75° the second at λʘ = 41.25° which corresponds with the very high threshold orbits of the concentration we found. The other sub-maxima that appear for the very low (blue) and low (green) threshold should be regarded as likely statistical fluctuations within the scatter of these dispersed orbits. The visible maxima in the high (red) and very high (yellow) threshold class are real.
The time span of 30° < λʘ < 46° that we selected to locate the concentration of HVI orbits is shorter than the actual shower duration. To have an idea of the actual shower duration we use the high threshold orbit from Table 2 to search for similar orbits among the 1101924 orbits we have in our dataset. The lookup with this reference orbit results in as many as 2796 possible HVI orbits that fit the minimal discrimination criteria. Determining an activity period for a shower depends on the tolerance we define for outliers. How many pure chance similarities do we allow? Most authors are not transparent at all about their criteria to identify orbits as shower orbits. Some use very low threshold similarity criteria; others use more rigid criteria. In Table 3 we list the intervals in solar longitude between which HVI orbits were detected for the different threshold classes.
Table 3 – The solar longitudes for the first and last HVI orbit detected for each class of similarity threshold.
|Similarity threshold||Begin λʘ||End λʘ||Orbits|
The HVI radiant
The number of orbits for each class is much higher than in our selection listed in Table 2, not only because of the much longer activity period, but also because of the much larger radiant size. In Table 4 we list the radiant sizes obtained when all orbits are taken into account.
Table 4 – The radiant area in Sun centered ecliptic coordinates for HVI orbits, for each class of similarity threshold.
|Similarity threshold||λ – λʘ (°)||β (°)||Orbits|
Among the extra orbits we find plenty which were previously identified as alpha Virginids (AVB#021), sigma Leonids (SLE#136) and April theta Virginids (ATV#730). Most of the ‘early’ HVI orbits were listed as AVB and SLE which seem to be different components of one and the same meteor stream complex active for a period of at least eight weeks. This confirms what Masahiro Koseki mentioned before (Koseki, 2020).
The search on the complete dataset also revealed a number of similar orbits registered late August till mid-September, around λʘ = 155° (August 28) from a radiant near αg = 182° and δg = +23°, in the region of Coma Berenices. It is rather diffuse and not worth further investigation unless more activity from this region would be confirmed.
The selection we used to locate the optimal reference orbit was limited to the interval 156° < λ – λʘ < 176° and –7° < β < +5°. The reason for this limitation in size is that a larger sampling area would include mainly outliers while the purpose is to locate a concentration of orbits to compute a mean orbit. Low velocity meteor streams near the ecliptic with short period orbits typically produce very large diffuse radiants. Using D-criteria on this type of orbits requires caution and therefore it is recommended to focus on the core of the shower. Figure 4 shows a compact concentration of radiants of high and very high threshold orbits. Figure 5 shows this for the inclination i in function of the length of perihelion Π.
Figure 6 shows the variation in geocentric velocity within the radiant. HVI meteors at left appeared slower than those at right. In Figure 7 we see the increase in vg in function of λ – λʘ. What we see is the effect of the Earth moving on its orbit around the Sun in the direction of the Apex. Figure 6 can be seen as a close up of the HVI position in Figure 8, with the x-axis in Figure 6 inversed.
When we consider the spread in the heliocentric velocity vh in Figure 9, no pattern in the velocity can be seen, the scatter seen is due to the error margin on the computed heliocentric velocities. No pattern appears in the plot of inclination i in function of the length of perihelion because of the error margins combined with the very low inclination close to zero (Figure 10). We also find a decrease of 0.13 km/s in the geocentric velocity per degree λʘ. HVI meteors appear to be faster at the begin of the activity than at the end (Figure 11).
Checking the orbital elements in function of solar longitude, we see that eccentricity e and the semi-major axis a remain stable. No trend can be derived within the error margins and scatter. The inclination i displays variations close to zero. The perihelion distance q shows an increasing trend (Figure 12) while the length of perihelion Π (= Ω + ω) indicates some increase but within a lot of scatter (Figure 13).
HVI component of a complex system?
Looking at the complete activity period we established for our reference orbit, searching our dataset with 1101924 orbits, we found many more possible HVI orbits. The distribution with the number of possible HVI orbits per degree of solar longitude is plotted in Figure 14. The sample we used to locate the HVI concentration was limited at 30° < λʘ < 46°. Figure 14 suggests there is HVI-related activity with another sub-maximum going on before our sampling window. However, caution is required with D-criteria applied with this type of orbits. To reveal the details of this earlier HVI related activity, a separate analysis should be made on the observing window 15° < λʘ < 35° to find the best fitting mean orbit for the concentration in this region. This study may answer the suggestion by Masahiro Koseki that some other known sources are related to this activity.
4 2020 activity by SonotaCo Network
The 2020 enhanced activity of the h Virginids has been observed by SonotaCo Network. In total 38 HVI orbits were identified. The mean orbits are listed in Table 5.
A simple average has been taken for the orbital parameters, 27 orbits had their radiants north of the ecliptic, 11 orbits south. This way the ascending node Ω and thus also the argument of perihelion ω switches 180°, therefore both groups were averaged separately. To compare the result with other networks, all SonotaCo Network orbits within the timelapse 31°< λʘ< 41° were used to calculate the mean orbit with the method of Jopek et al. (2006). The results agree well with the orbits obtained for previous years listed in Table 2.
Figure 15 shows clearly the radiant drift as well as the decrease in geocentric velocity vg, also found in all previous year’s orbit data (see Figure 11). No trend can be established for the absolute magnitude in function of time.
Figure 16 shows the radiant positions for the orbits in the nights of enhanced activity.
Figure 17 shows the orbits north and south of the ecliptic. Note the aphelia close to the orbit of Jupiter which has influence on the evolution of this shower. Mr. Yasuo Shiba (http://sonotaco.jp/forum/viewtopic.php?t=4580) suggests a possible 6-year periodicity, given that the average orbital period is 5.3 years and a 2:1 resonance with Jupiter’s revolution is valid.
Table 5 – The averaged HVI orbits north and south of the ecliptic compared with the mean orbit computed according to Jopek et al. (2006).
β > 0° (A)
β < 0° (B)
31°< λʘ< 41°
|vg||18.7 km/s||21.2 km/s||18.8 km/s|
|a||2.90 AU||2.84 AU||2.86 AU|
|q||0.7445 AU||0.6527 AU||0.7397 AU|
5 2020 activity by CAMS BeNeLux
Also CAMS BeNeLux recorded the 2020 enhanced activity. 65 orbits were selected within the interval 30.9°< λʘ < 46.9°, 156°< λ – λʘ < 176°, –7° < β < +5° and 13 km/s < vg < 23 km/s. The reference orbit for the concentration of HVI orbits was computed with an iterative procedure for all five threshold classes of similarity.
The results are listed in Table 6 and are in good agreement. The radiant plot for Sun centered ecliptic coordinates has been plotted in Figure 18. The concentration of the very similar orbits is very well visible. Figure 19 shows the increase in geocentric velocity vg within the radiant in the direction of the apex in Sun centered ecliptic coordinates.
Using the high threshold class orbits we find a decrease in geocentric velocity with Δvg/Δλʘ = 0.18 km/s, slightly higher than found for previous years (0.13 km/s per degree solar longitude) (Figure 20).
Table 6 – The reference orbits obtained from the 2020 HVI orbits by CAMS BeNeLux, for each threshold class of similarity.
|Medium High||High||Very High|
|vg||19.0 km/s||18.9 km/s||18.9 km/s||18.9 km/s||19.0 km/s|
|a||2.90 AU||2.94 AU||2.97 AU||2.97 AU||2.90 AU|
|q||0.739 AU||0.742 AU||0.746 AU||0.745 AU||0.739 AU|
6 2020 activity by the Global Meteor Network
The Global Meteor Network online orbit data for April and May 2020 includes as many as 191 orbits identified as h Virginids (HVI#343). In first instance we computed the mean orbit based on the identification as listed. GMN had significant more HVI#343 orbits than CAMS BeNeLux or SonotaCo Network. A detailed look at the dataset revealed that the list contained duplicates, slightly different orbits obtained from different camera combinations. In some cases the same camera combination yielded two orbits based on the same meteor which was detected twice on one of the cameras. Therefore it was decided to redo the analyzis on the GMN dataset after first removing all duplicate orbits.
Having done the analysis on the 2007–2019 HVI-orbits and knowing that the HVI activity covers a longer period than previously assumed, we decided to search for HVI orbits in a longer period of time taking into account a larger radiant area. Selecting all orbits obtained in the time interval 20° < λʘ < 55°, from a radiant area with 150° < λ–λʘ < 180° and –15° < β < +15°, with 12 km/s < vg < 23 km/s, as many as 727 orbits were available within this selection. A mean orbit was obtained by an iterative method with 494 orbits fitting the low threshold similarity class, 371 the medium low class, 213 the medium high class and 151 the high threshold class. Before solar longitude λʘ = 32° (April 22), many of these orbits were identified as α-Virginids (AVB#021) characterized by a higher inclination i with more northern radiant positions and some as April theta Virginids (ATV#730) characterized by a slightly lower velocity and higher values for the perihelion distance q. All of these are most likely components of a greater complex.
The longer the activity period of a stream, the larger the spread on the orbits of particles which must be widely dispersed in order to encounter Earth during a long period of time. Short orbit meteor streams near the ecliptic with low velocity appear from a very large radiant area and are difficult to distinguish from the sporadic background. The use of D-criteria is tricky in these cases and require caution. In this particular case with orbits north and south of the ecliptic and the inclination close to zero, the error margins on the orbits cause huge differences on both the ascending node Ω as well as the argument of perihelion ω. With inclination i = 0°, the orbit plane lies in the ecliptic plane and the nodes cannot be determined. Considering a long activity period and large radiant area results in mainly many more low threshold orbits with a high risk for contamination with sporadics and other nearby sources. Such large spread is not desirable to establish a mean orbit.
The real concentration of orbits within the high threshold class appears in the time interval of 30° < λʘ < 45°. Therefore we repeat the analyzes a third time, considering only the orbits in this time interval, limiting the radiant area to 156°< λ – λʘ < 176°, –7° < β < +5°. 240 orbits are available within this interval. The mean orbit of these 240 orbits is taken to launch the iteration to find the best fitting mean orbit for the concentration of similar orbits. Already after 4 iterations the procedure converges at a best fitting mean orbit for the high threshold similarity class. Next, the mean orbits are calculated for each threshold class separately using the method of Jopek et al. (2006). The final orbits for the high and very high threshold similarity class are almost identical to those obtained for the longer activity period and larger radiant area dataset mentioned above. Also the mean orbits obtained during the first attempt on GMN data were about the same, so the duplicated orbits did not influence the result much.
The number of orbits identified as HVI members for each class of similarity are shown in Figure 21 counted for each degree in solar longitude. The enhanced activity lasted during four nights from solar longitude 37 till 40° included (April 27 till May 1). The D-criteria show most of these orbits are very similar to each other. The final mean orbits for each similarity class are listed in Table 7. The dataset which covered a longer activity period and larger radiant area produced mainly many more low and medium low threshold class orbits with more scatter on their mean orbits, but almost identical for the high and very high class orbits.
Table 7 – The mean HVI orbits for the different similarity classes obtained from the 2020 Global Meteor Network data.
|Medium High||High||Very High|
|vg||18.5 km/s||18.5 km/s||18.5 km/s||18.5 km/s||18.5 km/s|
|a||2.95 AU||2.82 AU||2.89 AU||2.92 AU||2.95 AU|
|q||0.755 AU||0.751 AU||0.752 AU||0.753 AU||0.755 AU|
The HVI radiants in Sun centered ecliptic coordinates are displayed in Figure 22. The low and medium low threshold orbits (blue and green) appear very dispersed. A compact radiant (yellow) appears as a very narrow concentration of almost identical orbits. The velocity variation within the radiant is displayed in Figure 23, see also Figures 6 and 19.
The decrease in geocentric velocity in function of time is obvious in Figure 24 with Δvg/Δλʘ = 0.14 km/s, a bit less than found for CAMS BeNeLux, but almost identical with the long term value for this shower.
The number of HVI orbits identified in the Global Meteor Network data allows to compare the 2020 data with the long term data for the change in orbital elements in function of time (Figure 25). We see a clear increase in perihelion distance q while the other elements, semi major axis a, eccentricity e and length of perihelion Π, show no relevant trend within the spread on the values.
This case study has proven that the only reference orbit list in the IAU working list of meteor showers is not representative for the h Virginids (HVI#343). A new reliable reference orbit for this shower has been obtained, based on long-term orbit data obtained by EDMOND, SonotaCo Network and CAMS. The orbits for the 2020 outburst of h Virginids have been analyzed independently for the three datasets available for Sonotaco Network in Japan, CAMS BeNeLux network and the Global Meteor Network. The resulting mean orbits for the three 2020 datasets agree very well with the long term reference orbit. The final orbits are compared in Table 8. These orbits have a TJ value typical for Jupiter family comets. The activity period of the shower covers 28°< λʘ < 51° but similar orbits are found beyond these dates, identified as α-Virginids (AVB#021) before the main HVI activity and as April theta Virginids (ATV#730) after this interval. The 2020 enhanced activity may fit in some 6-year periodicity as also 2008 the year of discovery had good numbers of HVI orbits.
Table 8 – The long term reference orbit for the HVI#343 meteor stream compared to the mean orbits for 2020 as obtained by three independent video camera networks.
The authors thank Denis Vida for providing the scripts to plot the velocity distribution with a color gradient and to compute the average orbit according to the method of Jopek et al. (2006).
We are very grateful to the volunteers who maintain the IAU working list of meteor showers (Jopek and Kaňuchová, 2014, 2017; Jopek and Jenniskens, 2011).
We used the data of the Global Meteor Network which is released under the CC BY 4.0 license. We thank the SonotaCo Network members in Japan who have been observing every night for more than 10 years, making it possible to consult their orbits. We thank the camera operators of the CAMS networks. And we thank the contributors to EDMOND, including: BOAM (Base des Observateurs Amateurs de Meteores, France), CEMeNt (Central European Meteor Network, cross-border network of Czech and Slovak amateur observers), CMN (Croatian Meteor Network or HrvatskaMeteorskaMreza, Croatia), FMA (Fachgruppe Meteorastronomie, Switzerland), HMN (HungarianMeteor Network or Magyar Hullocsillagok Egyesulet, Hungary), IMO VMN (IMO Video Meteor Network), MeteorsUA (Ukraine), IMTN (Italian amateur observers in Italian Meteor and TLE Network, Italy), NEMETODE (Network for Meteor Triangulation and Orbit Determination, United Kingdom), PFN (Polish Fireball Network or Pracownia Komet i Meteorow, PkiM, Poland), Stjerneskud (Danish all-sky fireball cameras network, Denmark), SVMN (Slovak Video Meteor Network, Slovakia), UKMON (UK Meteor Observation Network, United Kingdom).
Drummond J. D. (1981). “A test of comet and meteor shower associations”. Icarus, 45, 545–553.
Jenniskens P., Nénon Q., Albers J., Gural P. S., Haberman B., Holman D., Morales R., Grigsby B. J., Samuels D. and Johannink C. (2016). “The established meteor showers as observed by CAMS”. Icarus, 266, 331–354.
Jenniskens P., Baggaley J., Crumpton I., Aldous P., Pokorny P., Janches D., Gural P. S., Samuels D., Albers J., Howell A., Johannink C., Breukers M., Odeh M., Moskovitz N., Collison J., Ganju S. (2018). “A survey of southern hemisphere meteor showers”. Planetary Space Science, 154, 21–29.
Jenniskens P. (2020). “h Virginid meteor shower”. CBET 4760.
Jopek T. J. (1993). “Remarks on the meteor orbital similarity D-criterion”. Icarus, 106, 603–607.
Jopek T. J., Rudawska R. and Pretka-Ziomek H. (2006). “Calculation of the mean orbit of a meteoroid stream”. Monthly Notices of the Royal Astronomical Society, 371, 1367–1372.
Jopek T. J., Jenniskens P. M. (2011). “The Working Group on Meteor Showers Nomenclature: A History, Current Status and a Call for Contributions”. In, W.J. Cooke, D.E. Moser, B.F. Hardin, and D. Janches, editors, Meteoroids: The Smallest Solar System Bodies, Proceedings of the Meteoroids Conference, held in Breckenridge, Colorado, USA, May 24-28, 2010. Edited by, NASA/CP-2011-216469,
Jopek T. J. and Kaňuchová Z. (2014). “Current status of the IAU MDC meteor showers database”. In, Jopek T. J., Rietmeijer F. J. M., Watanabe J., and Williams I. P., editors, Proceedings of the Meteoroids 2013 Conference, Poznań, Poland, 26-30 August 2013. A.M. University Press, pages 353–364.
Jopek T. J. and Kaňuchová Z. (2017). “IAU Meteor Data Center-the shower database: A status report” Planetary and Space Science, 143, 3–6.
Masahiro K. (2020). “Confusions in IAUMDC Meteor Shower Database (SD)”. eMetN, 5, 93–111.
Molau S., Rendtel J. (2009). “A Comprehensive List of Meteor Showers Obtained from 10 Years of Observations with the IMO Video Meteor Network”. WGN, Journal of the International Meteor Organization, 37, 98–121.
Roggemans P., Johannink C. and Cambell-Burns P. (2019). “October Ursae Majorids (OCU#333)”. eMetN, 4, 55–64.
SonotaCo (2009). “A meteor shower catalog based on video observations in 2007-2008”. WGN, Journal of the International Meteor Organization, 37, 55–62.
Southworth R. R. and Hawkins G. S. (1963). “Statistics of meteor streams”. Smithson. Contrib. Astrophys., 7, 261–286.