New Results from Video Meteor Observations

On several occasions ([1],[2]) I have introduced our video system MOVIE and the way it works. I was able to show quite impressive meteor videos and described the used hardware. Finally first results from the analysis of video meteors were presented on last years IMC ([3]).
Since then we had some more successful video observation sessions and did a lot of new investigations with the video data. This January we observed the Quadrantids near Hannover (Germany) and recorded more than 100 meteors parallel to our visual observations on video tapes (among them a nice -4 mag Delta Cancrid in CMa) with MOVIE ([4]). In the following weeks I did the tremendous work of analysing the Quadrantid tapes as well as the Perseid video from August 11/12, 1993, which was still unanalysed up to that time. There were another 250 meteors on these tapes, recorded during 7 hours of observation in the Schwarzwald mountains (Germany). After finishing this (all video data are now stored in PosDat format and available for every interested observer from Visual Commission Director Rainer Arlt) I had the necessary data basis for interesting research work in different fields of meteor astronomy and did some first calculations. The results were presented at the annually meeting of the German Meteor Observers Society (AKM e.V.) in Kirchheim (March 1995), here follows a summary of the most interesting outcomes.

One of the main aims of our video work is the accurate determination of radiant positions connected with the search for sub-radiant structures. In Belogradchik I present a first radiant plot for the Perseids 1994. Unfortunately bad weather conditions last year lead us to record most meteors far away from the radiant in the summer triangle and only very few in other region such as Andromeda/Pegasus. So the resulting plot ([3]) showed only a longish, inaccurate maximum near the predicted radiant position. There was no reliable statement possible, whether or not the are faint structures in the shower radiant of the Perseids.
This year we planned to observe the Quadrantids at a distance of about 30° from the radiant, because we wanted to obtain precise double station video observations together with our Dutch friends from the NVWS. Unfortunately the weather stopped our observation with clouds early in the morning and a guy called Murphy did the rest of the job: We had to finish just at the time, when the second video team 20 km away restarted their observation after their sky became clear. Thus we again did not managed it to record double station meteors. In addition our mounting did not drive the video system, because it was too cold (-8° Celsius). This is why the radiant slowly rotated into the field of view, and we captured many short meteors around the radiant. Last but not least someone stumbled in the middle of the night over a power socket, which caused the lens heating stop to work. Even though the resulting ice layer on the lens became thicker and thicker with time, it finally was a quite successful observation. We found almost 80 Quadrantids on the video tapes and could produce a nice radiant plot for this shower. Figure 1 shows this plot using the tracing method of Rainer Arlt's RADIANT software, figure 2 contains the same 39 meteors using the intersection method. This means, that complete meteor trails are traced back to the radiant in the first image, whereas the second picture shows only all the intersection point between two distinct meteor tracings.

Figure 1 - radiant plot for 39 Quadrantids from January 3, 1995, using the tracing method

---

Figure 2 - radiant plot for the '95 Quadrantids using the intersection method

It is obvious, that the 'theoretic' position of the radiant (shown as a green circle with a diameter of 5°) given in the IMO publications is very good. Furthermore there is no sub-radiant structure visible, even though the plot is very accurate and could show such features. So the absence of distinct structures within the Quadrantids radiant at the level of about one degree is the main result of this analysis.
As usual I tried to produce a nice picture of the shower (figure 3), which looks quite different from the Perseid image I presented last year (figure 4) at the IMC.

Figure 3 - shower picture of the Quadrantids

---

Figure 4 - shower picture of the Perseids

The meteors near the radiant are very short, we even recorded two pointlike meteors, which did not move at all.
In addition to this image I produced a computer animation, that shows the meteors appearing and disappearing dynamically around the radiant of the Quadrantids. During a few seconds 18 meteors with different lengths, velocities and brightness are visible on the screen, which illustrates all the well known effects of meteor showers quite impressive. After I have converted this animation into a standard format, I will make it available to everybody interested in it via my WWW homepage or by other means.
The next interesting shower have been the Perseids. It took me several days to analyse all the meteors from their maximum night '93, but then I had a database with more than 300 shower meteors available. In contrary to last year almost all meteors in 1993 were recorded in the morning hours and came from the Andromeda/Pegasus region, so the data sets from both years complement one another very good. The accuracy of parts of the data is not as good as for the Quadrantids, because I used an earlier version of the analysis software last year. In return I had a factor of ten more meteors available for the radiant plot. Figure 5 gives the distribution of 228 Perseids around the radiant. Their mean distance from the radiant is obviously still quite large, the violet ring marks a distance of 100° from the centre. Figure 6 and 7 show the radiant plot for these meteors again using the tracing and intersection method of RADIANT.

Figure 5 - distribution of 228 Perseids around the radiant using video observations from August 1993 and 1994

---

Figure 6 - radiant plot for the Perseids using the tracing method

---

Figure 7 - radiant plot for the Perseids using the intersection method

The meteors scatter more around the radiant, so the resulting peak is not as sharp as for the Quadrantids. The mean position of the radiant fits again quite well with the data given in IMO's meteor shower list. There are some minor sub-radiant structures visible in the plot, but I do not believe in the significance of these irregularities. The positional accuracy of each single meteor was only 1-2° near the radiant and the distribution of the meteors is still not optimal. So these structures are most probably artefacts.
One more interesting fact is the good agreement in the radiant position using two different methods (tracings/intersections). It seemed to me, that especially for higher numbers of meteors the later method gives better results, but both of them are equivalent on a first glimpse. Only the radiant position obtained using the probability algorithm shows a bigger difference, which is a subject of further investigation.

Beside the determination of radiant positions also ZHR calculations are an interesting area for video observers as shown on the last IMC. Some strange effects like abnormal high meteor rates during twilight were found at the first analysis but not confirmed yet.
The determination of zenith rates for this years Quadrantids was especially complicated due to the mentioned 'frozen lens' and the resulting large drop of the systems limiting magnitude. Nevertheless Jürgen Rendtel and I could show ([5]) a good qualitative correspondence of visual and video rates near the maximum. There is for instance a narrow peak in both activity graphs at 23:15 UT, which lasted only about 20 minutes.

One of the most interesting topics for me is the search for meteor clusters. In a paper from 1992 ([6]) I had analysed our visual meteor observations from that year searching for cluster effects and found absolutely nothing. Even though we had a good data basis (several hundred meteors observed from three visual observers in six successive nights with a time accuracy of 1 second) due to our computer based observation ([7]), the distribution of the meteors matched exactly the one expected for randomly in space distributed particles.
Two month ago I repeated this calculation for our video observation of the Perseid maximum night '93. This time I had to apply a special transformation first, because the standard formulae works only for constant meteor activity. This was definitely not given a few hours before the sharp ZHR peak. Using the same time resolution as for the visual analysis in 1992 I found again no evidence for any type of clustering ([8]). Figure 8 gives an idea, how good the theory for randomly distributed particles (exponential distribution) fits the observational results. The distance between two successive meteors is plotted on the x-axis, added up in intervals of 20 seconds length. The y-axis gives the percentage of each class compared to the whole number of 337 pairs of meteors.

Figure 8 - meteor cluster analysis with 20s intervals for 338 meteors on August 11/12, 1993

My suspicion now was, that clustering appears only on very short time scales (1-2 seconds), which might be smeared out in the 20 second intervals given above. So I did another calculation with an interval length of only one second, that is presented in figure 9. Here I used cumulative intervals to have more meteors in each class and get better statistics by it.

Figure 9 - meteor cluster analysis with cumulative 1s intervals for 338 meteors on August 11/12, 1993

Again one can clearly see, that there is almost no difference between (clusterless) theory and video observation. But if you look close enough to the very first intervals (up to a time distance of 12 seconds) you will see, that the observation shows always slightly more meteor pairs than expected! To make this clearer I added another graph to the diagram, which represents the relative differences between both values. We find a surplus of 57% in the first (meteor distances less or equal 1 second) and more positive differences in the following intervals. This implies, that there really might be some type of clustering of meteors at the Perseid maximum.
Looking at the statistics we should not forget, that this is a weak first clue: 57% surplus simply means, that we observed 11 pairs of meteors instead of 7 expected from theory. 30% surplus of meteor pairs with less or equal than 3 seconds distance stands for 21 pairs instead of 16,1. Furthermore I had to apply the mentioned special transformation for variable ZHR, which makes the results even more inaccurate. At least we have here for the first time a quantitative indication for a cluster effect at a low level of about 1.5%. This number results from additional computation to find the best fit between observation and theory and should be regarded as a dimension for the phenomenon only. All values between about 0.5% and 3% are thinkable too, because the calculations were quite unstable in relation to the used interval length and model.
Again the data basis is still not complete enough to give more precise statements at this time.

There are other interesting effects, which have to be confirmed in the future too. In Belogradchik I showed, that visual observers regularly underestimate meteor brightness by about 1 magnitude ([3]). A possible explanation is, that we estimate the brightness from the impression of the whole meteor trail, whereas video systems determine it at a scan rate of 25 measurements per seconds and therefore really obtain the absolute maximum brightness. This effect was dominant during the latest analysis of the Quadrantids and Perseids too, but I have not done another quantitative calculation yet.
Our latest video data provide a good basis for statistical analysis of meteor light curves along their path. This work remains for the next months as well.

To sum it up it can be said, that video systems have again proved to be very powerful tools in meteor observation. They provide large amounts of accurate data as a basis for many different investigation.

Reference