Forward scatter radio observations, Leonids, 2006

The carriers of Band 1 TV transmitters in Western Europe provide a convenient source of radio illumination of the atmosphere, allowing the use of modest receiving equipment to effectively observe the incoming meteors. These observations from northern England make use of transmitters in Norway and Belgium. See [1] for a good introduction to the physics of forward scatter meteor observation.

This year, Europe was well placed to observe the Earth's encounter with a dust trail left behind by the 1932 passage of comet 55P/Tempel-Tuttle. A peak of activity was predicted (David Asher [2]) for Nov 19th 04:45 UT geocentric, with a +4 minute topocentric correction for the UK, and a ZHR of around 120. Some very graphic and detailed computer predictions of the 2006 encounter with 55P/Tempel-Tuttle dust trails are available at [3].


A simple wire turnstile antenna, tuned to 55Mhz and placed about 1/4 wavelength above flat ground, points towards the zenith to provide omnidirectional coverage of the sky. An AR8200 receiver in USB mode delivers an audio signal to a PC soundcard for recording to disk. A receiver frequency of 55.2493Mhz places several distant TV carriers in the passband, two of which are used for these observations. This arrangement collects around a dozen meteor 'pings' per minute under background conditions, as well as reflections from aircraft and satellites.


Previous attempts to count individual pings during Leonid showers have not been successful due to the large number of strong overlapping reflections from overdense meteor trails, which can persist for several minutes. The simplest way to gauge activity seems to be to report the average signal level received in a narrow bandwidth centered on the carrier frequencies. Software analyses the data stream from the soundcard to determine the relative power in each carrier, using 50Hz-wide filters and applies a normalisation using a nearby empty band to compensate for receiver AGC action and broadband noise.

Unfortunately, it is not feasible to provide a calibration capable of converting average signal strength into the ZHR estimates which are of most interest to astronomers studying the dynamics and structure of comet dust trails. One problem is that the signal level can be dominated by a small number of long lasting overdense returns. Another difficulty is that radio and visual observations are likely to be sensitive to different particle sizes.

Daily Results, Nov 15th to 21th

The following daily charts use an integration time of 2 minutes and combine the results from the two most reliable TV carriers.

Leonid activity is present for several days, characterised by short bursts of fairly strong activity set against an elevated background consisting of a large number of faint pings.

Raw data files are available - email.

Visual Meteors

A total of eight visual meteors were seen between 04:15 and 04:45 on the 19th, through about 50% cloud cover with LM around 5 or 6. Five occurred in the first 15 minutes, two of which were coincident with radio echoes.

1932 Trail Encounter

Below is a closer look at the morning of the 19th.

There is a complete absence of any peak of activity near the predicted time of 04:49, although a period of increased activity begins at about 04:55 and continues for about an hour. An earlier burst of activity lasting 15 minutes was centered at 04:18 and coincident visual observation suggested a temporary ZHR of perhaps 70 or more.


It was disappointing not to see a clear peak of activity resulting from the 1932 trail. This was most likely due to the forward scatter system responding to a rather smaller particle size than the visual meteors for which the predictions were made (smaller particles are much more susceptible to radiation pressure). Early results (for example [5]) suggest that visual observers had more success. The recording process worked well, with no interference or breakdowns, but the signal analysis needs to be improved. The 'spikey' character of the charts suggests that the signal strength average is probably quite a non-linear function of meteor activity. Some ingenuity will be needed to extract a proportional count of meteor pings from beneath the strong overdense returns.

Example spectrograms

In these examples, time runs left to right, frequency from bottom to top.

Two meteors arriving about a second apart. The slanting lines are the Doppler-shifted reflections from the rapidly decelerating head of the meteor, which in this example last for about 2 seconds each. The broken horizontal lines are rapidly decaying reflections from the almost stationary underdense ionisation trails. The dark vertical bands are caused by the receiver AGC responding to the strong trail returns.

Orbit number 45733 of the International Space Station appears in this spectrogram as the broken slanting lines and the faint curved line. The horizontal lines are drifting overdense trails, here showing up five distinct TV carriers. Fresnel oscillations of the trail returns are clearly visible by the way they modulate the background noise via receiver AGC action. This spectrogram spans 50 seconds.

An example sound sample is available from the 2002 Leonids, which illustrates a typical mix of meteor returns during shower activity [4].


[1] Wislez J.-M.,Proceedings of the International Meteor Conference, Brandenburg 14-17 September 1995, "Forward Scattering of Radio Waves off Meteor Trails".
[2] Armagh Observatory The Leonid Meteors 2006
[3] Jeremie Vaubaillon, IMCCE, Computer Modelling
[4] MP3 sample (215kB, 70 seconds) from Leonids 2002, leo02-236.960.mp3.
[5] Visual observations, IMO

Paul Nicholson