Radio Meteor Forward Scatter Observations during Leonids 2000

Notes on an amateur observation. Please bear in mind that these results are not peer reviewed in any way. Last modified, 12th Nov 2001.
Frequency: 55.25 MHz
Transmitter: TV signal from Liege Bol d'air, Belgium.
Receiver: Todmorden, UK. 53:42:00 N, 2:04:00 W. 
Path:  Approx 500kM.
Recording: from 16/11/00 08:00 UT to 20/11/00 13:52 UT.

Reception Method

Zenith pointing wire turnstile with ground as reflector. AR8200 scanner receiver, upper sideband mode, 55.2493 MHz, audio output to PC soundcard. Data recorded in unsigned bytes at 8018.609 samples/second. Total 2.74 Gbytes.

Signal Characteristics

The chosen frequency places at least six continuous transmissions in the receiver passband, the two strongest being TV carriers at around 370Hz and 1650Hz. There are two other weaker carriers within 20 Hz of the strong 370Hz carrier, and two further signals at around 900Hz and 1100Hz, the latter spread into a spectrum of 12Hz sidebands covering about 100Hz. Pings are received independently on all these signals, although only the two strongest are used. There is also a wideband FM transmission within the passband whose pings are occasionally heard but this signal is neither used nor troublesome. Typical daytime background conditions are in Spectrogram 16/11 17:48.

Mean Relative Signal Strength

The plot shows the combined mean signal strength of the two strongest TV carriers, each received in a 150Hz bandwidth and normalised by the mean level from a 100Hz wide channel centered at 2110Hz. Each sample point represents a 600 second integration. The approximate positions of Asher-McNaught dust trail predictions [1][2] are marked.

All times below are UT +/- 5 mins. Using the strength 3 level as a threshold, the 1866 peak runs from 18.275 (06:36), to 18.335 (08:02), placing the center at 18.305 (07:19). The peak signal occurs at 18.281 (06:45).

The 1866 peak of the 1650Hz carrier alone (See 1650 Hz signal strength) is better defined and runs from 18.276 (06:38), to 18.336 (08:04), center at 18.306 (07:21). The 1650Hz peak signal is at 18.330 (07:55).

Count of Head Echoes

Software was arranged to recognise and count meteor head echoes from the 370Hz carrier, recognised by the rapidly falling doppler shift captured in the range 500Hz to 760Hz. Head detector was muted by a noise pulse detector monitoring a 120Hz bandwidth centered on 200Hz. The above plot shows the hourly rate. The peaks appearing between 19.74 and 19.95 are caused by local interference at the receiver.

Times of the four strongest peak head counts, UT +/- 5 mins.

   17.288 (17th 06:55)
   17.323 (17th 07:45)
   18.260 (18th 06:15)
   18.309 (18th 07:25) strongest

Further Processing

In an attempt to better distinguish the location of the signal strength peaks, a gaussian shaped sliding window was passed across the raw (combined carriers) signal strength, recording the gaussian weighted signal strength within the window at 1 minute steps of the window center. The peak locations resulting from the smoothed signal strength vary a little with the choice of sigma, as shown below - all times UT, +/- 1 minute.

90 mins 17.3403 (17th 08:10) 18.3004 (18th 07:13) Plot
30 mins 17.3205 (17th 07:42) 18.3056 (18th 07:20) Plot
20 mins 17.3250 (17th 07:48) 18.3035 (18th 07:17) Plot

One final chart was done using a sigma of 8 minutes (Plot). At this resolution the two main peaks each begin to split into three sub-peaks. The times of the six peaks are

1932:17.2958 (07:06); 17.3285 (07:53); 17.3757 (09:01)
1866:18.2928 (07:02); 18.3056 (07:20); 18.3306 (07:56)


The steady rate of soft background pings turns into a bedlam of overlapping overdense returns during the shower, rendering useless any attempts to count individual pings or bursts. Resorting to a plot of signal strength reveals two solid peaks but the timing of these, and in particular, the location of the maxima, are ambiguous. The usual problems of calibrating the intensity prevent comparison with predicted ZHR amplitudes. An attempt was made to recognise and count the head echoes in the hope that these might remain countable throughout the event. This does seem to break up the peak of 18/11 into two distinct and fairly well defined maxima but further work needs to be done to ensure that these are not artifacts of the head counting process.

The signal level of the two channels combined is plotted below along with the predicted positions of the 55P/Tempel-Tuttle dust trails - based on a diagram by David Asher, Armagh Observatory;

We get a clear impression of the strong peak coinciding in both position and width with the 1866 dust trail, and this peak is preceeded by the shallower, broader peak of the 1733 trail. This latter is not really discernable as a separate peak, appearing instead as an extended introduction to the main peak. The 1932 trail from the previous morning also coincides nicely, apparently with a slightly extended tail.

Despite the plentiful supply of echoes, the location of the peaks remains open to interpretation, which demonstrates the need to combine results from several reception sites before a reliable picture of the meteor shower can be obtained. The sub-peaks apparent in some of the data are most likely to be due to the response of the signal averaging to just a few long overdense echoes, although a comparison with other radio meteor records would be interesting.


Signal analysis software in C, under Linux, built on Phil Karn's FFT libraries [3]. XForms library [4] for the display. Source available if anyone wants it. See screenshot.


[1] Armagh Observatory, The Leonid Meteors 2000.
[2] Astronomical Society of Australia, The Leonid Meteor Shower
[3] Phil Karn's FFT library, available from Seti League, Karn FFT
[4] XForms library, by Zhao and Overmars, XForms Home page.

Paul Nicholson