Go to Armagh Observatory Leonid page.

Rob McNaught and David Asher's final comments before the 1999 Leonid meteor shower, as shown below, are being sent to the IMO-News e-mailing list, after which the authors are setting off on their travels to observe the Leonids. The authors will reappear after the Leonid shower if the predictions do not turn out to be substantially wrong.

- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -

Note to readers:  if you do not have access to reference [1], please skip
sections that are incomprehensible without it.  Things such as the Table 3
predictions should be clear by themselves.

                                                     Last modified 1999 Nov 9


R.H. McNaught and D.J. Asher


In WGN 27, 85-102 (1999), details of the Earth's encounters with Leonid dust
trails were presented.  Moderately close encounters will lead to substantial
meteor outbursts during the 1999 and perhaps the 2000 Leonids, while even
closer encounters will produce storm level activity in 2001 and 2002.  Here
we summarise the predictions of the dust trail model for 1999 and the
following few years, for the last time before observations of the 1999
Leonids afford the first test of these predictions.  The updated analysis
here is a little more comprehensive than presented previously.


1. Introduction

The highest ZHR Leonid storms of the 19th and 20th centuries, as well as many
other sharp outbursts, have occurred when the Earth encountered a young dust
trail within the Leonid stream.  Such a trail is generated every 33 years or
so when Comet 55P/Tempel-Tuttle returns to perihelion.  Each trail
progressively lengthens while remaining narrow and dense (density dilution
being due to lengthening alone, not broadening) until it is scattered into
the Leonid background after a few centuries.  Readers should refer to [1] for
further details.

The predictive power of the dust trail theory is demonstrated by the
following facts.

   Storms and outbursts over the past 200 years that correspond to the
   Earth's encounter with a given young dust trail have the calculated peak
   times (i.e., based on the centre of the Earth reaching the calculated
   nodal longitude of the trail) and the observed peak times matching to
   10 minutes or less [1], in cases where the observed peak is known to
   better than that accuracy.

   A topocentric correction improves the match even further [2].  Data from
   the past 200 years now indicate that close encounters are predictable with
   an uncertainty of 5 minutes.

   The peak time of the 1998 Draconid outburst [3] was predicted equally
   accurately by Reznikov [4] using the same form of dust trail

   Moreover, Leonid timings relating to what we term young trail encounters
   have been independently confirmed by the Russian group that includes
   Reznikov [5,6] and by Lyytinen [7].  See [6] for references by the same
   authors describing work on other streams.

2. Update

In [1], a desire to avoid excessive effort led us to make our calculations
comprehensive (covering the past 200 years) only for trails 6 or less
revolutions old; three encounters with slightly older trails were considered
as special cases.  Also an error in the calculations caused very small
inaccuracies (less than 0.0001 AU) in the determination of trails' nodal

Now we provide an updated list of encounters (covering the next few years
only, these being of greatest interest for meteor observers) for trails up to
9 revolutions old.  Changes to the ZHR fit due to the error are not
substantial but a new fit is done.

Table 1 shows the data for past trail encounters.  The only corrections to
the entries in Tables 2 and 3 of [1] are in r_E-r_D.  Reference [1] can be
consulted for more details, but in summary, the strength of an outburst is
affected by Delta a_0 (which effectively corresponds to the ejection
velocity, this in turn being related to the mass distribution),
r_E-r_D (the miss distance of the Earth from the trail node) and f_M
(which measures the change in density due to trail lengthening).  Table 2
is for encounters over the next few years (cf. Table 5 of [1]).  Roughly,
f_M is expected to be inversely related to the age of the trail, but for
example, the value for the part of the 9-rev trail that is encountered
in 2001 shows that gravitational perturbations can cause deviations from
this simplistic relationship, after a few revolutions.

Table 1 - Past trail encounters
                                                Observed  Calculated
Year  Trail   Node   Delta a_0  r_E-r_D   f_M    ZHR/f_M   ZHR/f_M
             (J2000)    (AU)      (AU)
1966  2-rev  235.158   +0.168  -0.00013   0.52   170,000   100,000
1833  1-rev  233.184   +0.174  -0.00021   0.95    63,000    76,000
1866  4-rev  233.333   +0.059  -0.00029   0.37    22,000    22,000
1867  1-rev  233.420   +0.373  -0.00014   1.00     4,500     4,600
1869  3-rev  233.536   +0.320  -0.00047   0.44     2,300     2,200
1969  1-rev  235.272   +0.934  -0.00004   0.95         -         -

Table 2 - Future trail encounters

Year  Trail   Node   Delta a_0  r_E-r_D   f_M
             (J2000)    (AU)      (AU)
1999  3-rev  235.291   +0.138  -0.00066   0.38
2000  8-rev  236.103   +0.064  +0.00076   0.27
2000  4-rev  236.276   +0.114  +0.00077   0.13
2001  7-rev  236.114   +0.081  -0.00043  ~0.14
2001  9-rev  236.429   +0.041  +0.00015   0.43
2001  4-rev  236.463   +0.142  +0.00022   0.13
2002  7-rev  236.610   +0.113  -0.00015   0.13
2002  4-rev  236.888   +0.172  -0.00005   0.15
2006  2-rev  236.615   +0.961  -0.00009   0.53

A model in which ZHR/f_M (ZHR being the observed peak ZHR in past encounters)
is fitted as a function of Delta a_0 and r_E-r_D, as described in [1], is
done, and applied to the future years.  The five storm years in Table 1
(1969 excluded) are used in the fit, to interpolate ZHR estimates for
1999-2002.  It is inappropriate to apply exactly the same model to very
different values of Delta a_0 and so 1969 alone is used to predict 2006 alone.
The predictions are in Table 3 (cf. Table 6 of [1]).  Only the fit centred at
r_D (cf. Tables 4 and 9 of [1]) is given, the formal uncertainty in the fit
being 20%.  Whilst the overall fit is reasonable, there is now a
discrepancy between the calculated ZHR values for 1833 and 1966.  For 1966
the calculated ZHR is 53,000.  Despite the uncertainty in the observed ZHR in
1966, it is probably one of the more reliable data points used in the fit and
1833 the least reliable.  In 1999, the formal ZHR prediction is 500 and
this appears fairly robust, regardless of whether the 1833, 1966 or both are
used in the fit, but values 200 < ZHR < 2000 give a reasonable fit.

Table 3 - Predictions

Time (UT)           Trail  Estimated  Moon   Visible from
                              ZHR     age
1999 Nov 18, 02:08  3-rev      500     10  Africa, Europe
2000 Nov 18, 03:44  8-rev       30?    22  W. Africa, W. Europe, NE S. America
2000 Nov 18, 07:51  4-rev       20?    22  N. America, Central America &
                                              NW S. America
2001 Nov 18, 10:01  7-rev    1,500?     3  N. & Central America
2001 Nov 18, 17:31  9-rev   15,000      3  Australia, E. Asia
2001 Nov 18, 18:19  4-rev   15,000      3  W. Australia, E., SE & Central Asia
2002 Nov 19, 04:00  7-rev   15,000     15  W. Africa, W. Europe, N. Canada,
                                              NE S. America
2002 Nov 19, 10:36  4-rev   25,000     15  N. America
2006 Nov 19, 04:45  2-rev      100     28  W. Europe, W. Africa

Figures 1 and 2 are the visibility maps that are not in [8].  Note that in
these figures, the Moon's phase is displayed as seen from the southern

Figure 1
   (See http://www.atnf.csiro.au/asa_www/images/2001csl.gif [15Kb])

Figure 2
   (See http://www.atnf.csiro.au/asa_www/images/2002bsl.gif [15Kb])

3. Discussion

Revised values of the nodal distance required a reassessment of the ZHR
predictions from the dust trail density model.  Overall, the rates have not
changed substantially, although the uncertainty in the fit has increased.
For 1999, the predicted ZHR for the 3-rev dust trail encounter is probably of
the order of 500.  This value requires some elaboration.  Activity from this
dust trail will be additional to background activity, which could itself
have a ZHR in the hundreds.  Thus, the observed ZHR at the time of the
peak will probably lie in the range 500-1000, if the dust trail contributes
a ZHR of 500.  This value is entirely consistent with past data,
given that there are uncertainties in the past peak ZHRs used in the
fit, and different fits can be done (cf. [1]) centred on slightly different
values of r_D.

Given that some older trails have been demonstrated to be capable of
delivering high rates (e.g. the 9-rev trail of high f_M in 2001), it will be
necessary to check that for the years of the storm data used in the ZHR fit,
no additional old dust trails were contaminating the ZHR.  In future
analyses, we shall also remove the background component from the peak ZHRs,
to more truly represent the contribution of the dust trails alone.

One interesting change to the ZHR fit, is that the predicted ZHR in 1801
from the 2-rev trail is now 300.  This is much smaller than our original
predictions that suggested a minor storm had occurred.  Thus, this potential
anomaly of an unobserved storm in the last 200 years over western Europe, is
no longer a problem.  A short lived peak ZHR of around 300 would probably not
have attracted much attention in those years.  However, we are aware of no
data that refute the possibility of a storm in that year.

There are no additional encounters with trails up to 19 revolutions old in
1999; therefore, unpredicted high activity is unlikely.  There appear to be
no other encounters of significance in the following years up to this age,
although in 2001, an encounter with a disrupted 10-rev trail of uncertain
density could produce a peak ZHR of around 1,000 on Nov 18, 18:01 UT.
Although the disrupted nature of this section of the 10-rev trail makes
this time unreliable (pending more detailed simulations), it appears to be
during the 48 minute gap between the stronger encounters.  These three
encounters in close succession, with no interference from the Moon, suggest
the highest observable rates will occur in 2001, although the ZHR in 2002 is
likely to be higher.

For the 3-rev trail encounter in 1999, the time of maximum is predicted to be
at Nov 18 02:08 UT in the Mediterranean region, with an uncertainty of around
5 minutes.  The time of maximum is dependent on location [2], with the peak
predicted at 01:58 in South Africa and 02:14 in northern Scandanavia.  The
dust trail model does not make any prediction about the time or intensity of
the background activity maximum, but in the past 200 years the highest
Leonid rates outside young dust trail encounters, can approach a ZHR of

Further information is available in [9].


[1]  R.H. McNaught, D.J. Asher, `Leonid dust trails and meteor storms.'  WGN
     27, 1999, pp. 85-102.

[2]  R.H. McNaught, D.J. Asher, `Variation of Leonid maximum times with
     location of observer.'  Meteorit. Planet. Sci. 34, 1999, pp. 975-978.

[3]  R. Arlt, `Bulletin 13 of the International Leonid Watch: The 1998 Leonid
     meteor shower.'  WGN 26, 1998, pp. 239-248.

[4]  E.A. Reznikov, `The Giacobini-Zinner Comet and Giacobinid meteor
     stream.'  Trudy Kazan. Gor. Astron. Obs. 53, 1993, pp. 80-101 (in
     Russian).  See also IMO-News mailing list, 1998 Sept 9.

[5]  E.D. Kondrat'eva, E.A. Reznikov, `Comet Tempel-Tuttle and the Leonid
     meteor swarm.'  Sol. Syst. Res. 19, 1985, pp. 96-101.

[6]  E.D. Kondrat'eva, I.N. Murav'eva, E.A. Reznikov, `On the forthcoming
     return of the Leonid meteoric swarm.'  Sol. Syst. Res. 31, 1997,
     pp. 489-492.

[7]  E. Lyytinen, `Leonid predictions for the years 1999-2007 with the
     satellite model of comets.'  Meta Res. Bull. 8, 1999, pp. 33-40.

[8]  R.H. McNaught, `Visibility of Leonid showers in 1999-2006 and 2034.'
     WGN 27, 1999, pp. 164-171.

[9]  Armagh Observatory Leonid WWW pages are http://www.arm.ac.uk/leonid/
     and general notes for the public are at


DJA thanks Esko Lyytinen for extremely valuable discussions on this work.

Authors' addresses

Robert H. McNaught, Siding Spring Observatory, Coonabarabran,
NSW 2357, Australia (rmn@aaocbn.aao.gov.au)

David Asher, Armagh Observatory, College Hill,
Armagh, BT61 9DG, N. Ireland, UK (dja@star.arm.ac.uk)

Last Revised: 19th January 2000
WWW contact: webmaster@star.arm.ac.uk
Go to HOME Page