International Meteor Organization (IMO)

A First Global Analysis of the 1994 Perseids

Jürgen Rendtel

Published in WGN, the Journal of IMO 22:6, p. 205 (1994)

Abstract: This preliminary analysis is based on more than 18000 Perseids recorded by more than 80 observers covering over 750 hours of effective observing time. The high Perseid peak recurred in 1994 at sol=139.595°±0.007° corresponding to 11h UT on August 12, and was observed from various sites in North America. The population index r was lower during this peak period (rP=1.8) than during the regular Perseid maximum around sol~140° (rM=2.1). The minimum of r coincides with the activity peak. The maximum equivalent ZHR (based on 10 minute counts) of the peak was EZHR=250±45. This is lower than in 1991 and 1993. The regular maximum reached a level of ZHR~90 which is also lower than in 1991 and 1993.

Introduction

After the surprising occurrence of a strong ZHR peak during the 1991 Perseids and the extreme expectations during the 1993 return, the 1994 recurrence became an almost routine observation. However, there were a number of open questions. Recent model calculations [1] indicated that the 1994 return of the sharp Perseid peak might be comparable to the 1993 event. The precise time of peak activity was not predictable, though. Considering the shifts observed in the previous years, the most probable time was August 12, 9h30m ±3h UT. Thus the most favorable region for observations was in the western part of the United States. Some observers in central and southern California suffered from clouds brought inland by a hurricane west of the coast. On the other hand, northern California and Oregon were cloudless during the respective period.

All previous Perseid outbursts were not optimally observed: in 1991 the observers, of course, were not prepared to record such a high activity. The Full Moon coinciding with the 1992 peak and the earlier than expected time of the peak resulted in data of poor quality only. In 1993, the peak occurred later than expected. Consequently, most European observers finished their observing during twilight, while the eastern parts of North America had poor skies and cloudy weather. The collection of data from many observers, however, resulted in a global analysis with reliable results [2]. However, data in the immediate vicinity of the peak, corresponding to the ascending branch and the descending branch of the higher activity, respectively, were from different observational sources, and hence suffer from problems caused by combining these data (e.g., the perception correction in the case of the Geminid maximum as discussed in [3]). In 1994 data sets of both the ascending and descending branch were obtained by the same observers under (almost) constant conditions with small corrections. These observations should enable us to determine more about both the peak itself and the observer's perception under these conditions (as already attempted in [4]). This will be the subject of another analysis.

The 1994 return

When observers in Europe finished their watches on August 12, 1994, at about 3h UT, the ZHR was at a level of 50. Observers in North America continued without a gap this year (due in part to better weather conditions), and thus we have a continuous series of data from sites where the radiant was high in the sky. Activity did not change until 7h UT, when a slight increase was noted. Between 9h30m and 10h UT, the ZHR exceeded 100 and the increase continued. This was reported by observers in the eastern part of the United States: as in 1993 in Europe, they witnessed the ultimate rise to maximum in the bright morning twilight. Just around 11h UT, the ZHR peaked at 250 followed by a remarkably steep decrease. About 40 minutes later, the ZHR fell to almost 100. As already pointed out, we do have counts of the entire profile obtained under constant circumstances, since the twilight permitted observations until 12h UT in California. This is most important for the determination of the population index r from the magnitude data.

The analysis was done from a (large) subset of the data sent in. Since it was intended to present this overview at the end of 1994, not all data are in the VMDB files yet. The sample used for this analysis contains data of almost 18000 Perseids observed in more than 750 hours effective observation time by more than 80 observers.

The population index

The most interesting part of the r-profile in Figure 1 is the peak period. Here, the value of r is lower than before and also lower compared with the average, or normal maximum around sol=140.0°-140.5°. This was not seen as clearly during the previous returns, as explained in the Introduction.

Figure 1 - Variation of the population index r for the period August 8-13, 1994. The profile around the peak at sol=139.595° is shown in more detail in Figure 2.

Figure 2 - Population index profile around the Perseid maximum. August 8 to 11 corresponds to the activity plateau of the Perseids at a ZHR level of roughly 30. In this period, r~2.1 with only slight variations. The values of r are averages of 1°-intervals in solar longitude, shifted by 0.5°.

During the ascending branch of the activity, the radiant elevation steadily increased for all observers. Although the change in the entry conditions should lead to an increase in r [6], we observe a net decrease and a significant minimum coinciding with the EZHR peak. Observations were sufficiently complete in 1994 to calculate reliable r-values for 1-hour interval lengths. Note the small error bars which suggest interpretation of the variations as real changes of r. The value of r of the new meteoroid population should be even lower as discussed in the text.

Figure 2 shows this phenomenon in more detail. Note that 0.04° corresponds to 1~hour.

Surprisingly, the increase of r with an increasing radiant elevation which was observed during the 1992 Quadrantid return [5] and later explained by Bellot [6] by varying entrance condition of the shower meteors with the radiant's elevation, did not occur during the 1994 Perseid peak. This was not to be derived from the previous data series because of their composition. But the 1994 Perseids were observed under quite similar conditions as the 1992 Quadrantids: the radiant elevation steadily increased until the end of the observations, i.e., until the peak was entirely passed. What we see is just the opposite tendency: the value of r decreases as the radiant approaches the zenith. If we consider the effects discussed by Bellot [6] acting here as well, we should assume an even lower minimum value of r.

Furthermore, the value of r during peak activity in 1994 is very close to the corresponding values of the returns in 1991 (rP=1.9, [7]) and 1993 (rP=1.8, [2]). The value of rM=2.1 found for the regular maximum is almost identical for all returns from 1991 to 1994 [2,7]. Again, we should consider that in 1994 the period around the peak is covered by more continuous and homogeneous data sets compared with the combined data obtained at various sites in previous returns. This does not devalue the earlier global data analyses, but the fewer corrections which are needed, the fewer systematic effects which are likely to be introduced. Thus we may now conclude more reliably that the particle population observed during the peak period is different from the average, older material observed outside this region. This difference should be even somewhat larger than determined by the difference in the population index r shown in Figures 1 and 2 because the total number of meteors (say, 200 per hour around the peak) also includes a portion of the regular material (approximately 40 per hour, or 20%).

The activity profile

There exists little experience with the observation and analysis of very high meteor activity. The measure we are used to is the ZHR. The peaks we observed for the Perseids recently last for less than an hour. Within one hour, the observer witnessed a substantial part of an activity profile, and if we use the term ZHR in its original sense, the rates were much lower. However, we simply use the quantity ZHR for shorter intervals, such as 15 or even 10 minutes, knowing that this measured rate did not last for an hour, but rather is an extrapolation of the instantaneous activity averaged over a smaller time interval. Immediately after the 1993 return (and again in 1994), there occurred notes about ZHRs of 700 or so. Most of these turned out to be derived from counts during intervals of less than 10 minutes. Of course, one may find very short intervals (such as single minutes) with 5-10 Perseids, and we could calculate a "ZHR" from these. It is obvious that these numbers have a completely different meaning. Firstly, they are based on a small sample, and, secondly, this sample is to a large extent determined by accidental factors (such as the appearance of a number of shower meteors quite close to each other in the sky). Furthermore, one bright fireball with a persistent train may attract the observer's attention and may let him miss other, faint shower meteors appearing in the same minute. Analyses of 1993 data indicates that count intervals of less than 10 minutes are the limit for accurate representation of the activity level observed in the Perseid peaks [4]. This may be somewhat shorter if the rates further increase. In order to make obvious that the rate is obtained from intervals which are significantly shorter than one hour, I propose use of the term "equivalent ZHR" (EZHR).

Figure 3 shows the ascending branch of Perseid activity as derived from the global data. As in 1993 [7], there is only little fluctuation in the rates. The peak itself at sol=139.595°±0.007° is well defined by 2 interval averages based on 10 independent count intervals containing a total of more than 700 Perseids. This time corresponds to 11h UT on August 12, 1994.

More detailed analyses of the immediate peak period will be carried out for individual returns. The maximum EZHR for 10 minute counts slightly exceeded 250. This seems to underline the steady decrease of the highest rates after the 1991 peak (Table 1).

Figure 3 - Ascending branch of activity and the peak at sol=139.595°±0.007°. The increase of the equivalent ZHR (EZHR) is continuous, and the peak is well pronounced. Here we shifted a sampling interval of 0.016° length (20 minutes) by 0.008°, using only the count intervals of <=15 minutes. More detailed analyses using only strictly identical count intervals are the subject of a further analysis. Here, we have to expect some smearing because the 15 minute count intervals are not identical and thus we in fact average over more than 20 minutes.

Table 1: Highest recorded equivalent ZHRs during the Perseid peaks of 1991 to 1994. The 1992 result should be left out from further conclusions because all data suffered severely from disturbance by the Full Moon.

Year New peak Regular maximum
rPEZHRrMZHR
19911.93502.1120
1992(2.1)(250)(2.1)(90)
19931.83002.0110
19941.82502.190

In Figure 4, we show the ZHR profile for the period which includes both the peak and the regular maximum. The ZHR of the regular maximum was just below 100. This seems a little lower than the average of the 1991 and 1993 returns when the ZHR of the maximum at sol~140° reached 120 [7] and 110 [2], respectively. Before we try to interpret this, we should include all available data and also consider the error bars of all rates found.

Figure 4 - The two Perseid maxima. The peak of fresh material at sol=139.595°±0.007° and the regular maximum at sol=140.1°-140.6° according to the preliminary analysis. For the regular maximum we used sampling periods of 0.2° length (5 hours), shifted by 0.1°. The five data points between sol=139.7° and sol=139.9° are averages of Japanese data submitted by Junichi Watanabe and added to the profile.

After the peak time was passed, the ZHR fell to about 80 before climbing to the regular maximum. This is exactly the ZHR that was previously observed when 139.6°<sol<139.9° before the new peak occurred (see, e.g., [8]). So we may state that the new peak represents additional activity superposed with the average rate we know from the regular Perseid returns.

Conclusions

Although this analysis is only based on part of the available data, the results can be expected to be quite close to the final values. This is particularly valid for the peak period. Further analyses will deal with details in the immediate vicinity of the peak.

A result which was not as prominent in the analyses of 1991 to 1993 Perseid data is the behavior of the population index r during the peak activity period. Despite the effects discussed by Bellot [6], the value of r decreased with increasing radiant elevation.

Combined with the fact that the observed particle population at the peak consists of approximately 20% "old" particles (not belonging to the peak caused by freshly released meteoroids), this means that the population index r of the new material is even lower than the value of r=1.8 derived from the observations. This was not clearly visible in the 1991 and 1993 data because of the composition of the data.

The uncertainties of all parameters obtained from light-disturbed observations, such as in 1992, underline that such data can only be used for deriving upper/lower limits of some parameters.

The shift of the new peak relative to the nodal crossing time for P/Swift-Tuttle was of the same order as observed between the previous returns. The analysis indicates that the 1995 peak should be expected to be weaker than in 1994, occurring on August 12 at 17h±3h UT.

Acknowledgments

I wish to thank all observers who sent in their data very soon after the Perseids. This enabled us to present a very first analysis at the International Meteor Conference in Belogradchik, Bulgaria [9]. The immediate contact with Peter Brown after the peak observations also enabled the IMO to report a reliable overview for the IAU Circulars. \newpar Special thanks go to Junichi Watanabe of Japan for sending some data from Japanese observers before we obtained the actual raw data for input into the VMDB. This allowed us to fill in the gap between the peak and the "traditional" maximum.

References

[1] I.P. Williams, Z. Wu, MNRAS 269, 1994, pp. 524-528

[2] J. Rendtel, WGN 21, 1993, pp. 235-239

[3] R. Arlt, J. Rendtel, WGN 22, 1994, pp. 167-172

[4] J. Rendtel, in Proc. Meteoroids Conference Bratislava, 1994 (in press)

[5] J. Rendtel, R. Koschack, R. Arlt, WGN 21, 1993, pp. 97-109

[6] L.R. Bellot, WGN 22, 1994, pp. 13-26

[7] J. Rendtel, R. Koschack, R. Arlt, WGN 21, 1993, pp. 152-167

[8] R. Koschack, P. Roggemans, WGN 19, 1991, pp. 87-98

[9] J. Rendtel, in Proc. International Meteor Conference, P. Roggemans, ed., Belogradchik, Bulgaria, 1994 (in press)

[IMO Homepage]

webmaster@imo.net; last change: November 10, 1997