WSPR and MH370, revisited: Some notes on Fading

It seems to be a never ending story: again Richard Godfrey and Dr. Robert Westphal, DJ4FF, go on a fool’s errand regarding “WSPR and MH370”. Against all physics and reason, they continue to try to prove that it is possible to detect aircraft – and now: missiles – based on the log data of weak WSPR signals documented every 110 seconds. Even WSPR developer and Nobel laureate Joe Taylor, K1JT, has relegated this to the realm of unscientific folly: “Anyone who does this doesn’t know what they’re doing.”

Although in the aviation press and also among HF experts the support of these charlanteries is rapidly dwindling, the duo nevertheless succeeds in promoting them successfully and with great fanfare, especially among radio amateurs and their media – “because they don’t know what they’re doing” (both parties). There a scientific discussion is suppressed so far, on the contrary: One feels reminded of the flickering will-o’-the-wisp in the windows of burning asylums.

The duo’s latest folly is titled “SpaceX Falcon Launch and WSPRnet Detection,” to be precise: not doing that. In this, they stir together ignorance of HF propagation and apparent lack of expertise of what WSPR can and cannot do into an unpalatable mush. Again, it is a matter of inferring aircraft from data in the WSPR log when the signal-to-noise ratio in the logs behaves “unusually”.

With another new approach, I would like to show that what is ordinary about HF propagation is precisely its unusual nature. For this, I studied the carrier of the Riyadh radio station on 15380kHz (two hops) over four consecutive days for two hours each. The transmitter radiates with 500kW transmit power at a HRS4/4/.5 with about 20dB gain towards 310° and is thus able to provide also for some aircraft scatter detectable by Doppler traces.

The first question is: How do the signals develop, day by day? The figure below shows the levels on the four days with a resolution of one second.

Same hours, different days: The received level of Riyadh transmitter.

The second question is: Is there any correlation of the levels? The figure below clearly says: Nope.

Riyadh’s levels: Zero correlation. See correlation coefficients in the top left corner.

But the WSPR log saves only the average SNRs in chunks of 110 seconds each. So, the third question is: How do those chunks develop day by day? The figure below gives an overview:

Here, the levels of the first figure have been averaged over chunks of 110 seconds each to simulate WSPR log data.

This also calls for a correlation matrix:

Correlation matrix of 110 seconds’ chunks. 8 of 12 combinations do show no correlation, whereas with four combinations we see a “weak correlation” – by mere chance.

The fourth question gets us right to the core: Can we see from the SNR data, by their “unusual change” some aircraft scatter?
To scrutinize this question, I calculated the level difference of one chunk to the next. The idea behind it: If aircraft scatter is detected, there will be an “unusual” change of the level, most probably to a higher one. See below for this figure:

Differences from chunk to chunk: What is “usual”, what is “unusual” propagation? A few critical points have been marked in red, see spectrograms below.

Now let’s check the spectrogram of the HF recording to see where there are some real aircraft scatter, and if they correlate with the peaks of above figure. Here, I did this only for the first day because the result is similar on all days:

The spectrogram shows a nice aircraft scatter (LSB) at 06:48UTC, where the chunks show nothing unusual: the averaged level reads about minus .1 dB!

So, “nothing heard”, as they say in DXpeditions … Let’s try it vice versa: we not the biggest peak around 08:28UTC. According to Godfrey and consorts this should be “unusual”, and hint towards an aircraft! But see the spectrum below: Nope!

When we note the biggest peak in the chunks’ levels, no aircraft scatter can be found!

I am convinced that these easy-to-understand, yet scientific accounts of the actual propagation conditions should convince even the simplest mind: The detection of aircraft etc. with WSPR log data is not possible.

Last but not least, the figure below shows how little aircraft scatter affects the carrier signal. Moreover, most of the Doppler signal is outside the 6 Hz bandwidth of a WSPR signal. And for fun, everyone can calculate that the Doppler signal is at least 40 dB below the carrier signal. If the carrier signal is -50 dBm and the Doppler signal -90 dBm, the latter would be raised by 0.0004 dBm. If, yes, if this Doppler trace fell directly on the carrier …

Doppler & carrier: A strong Doppler from a strong broadcaster’s carrier. If you follow the course of the carrier and that of the Doppler, you will see that correlations are purely random.

Surely some proponents of the thesis that nevertheless airplanes are detectable at great distances from WSPR log data (admittedly only on an earth as a disk, which they may believe in …) will not be further disturbed by the physics presented here. What again does not disturb me. As a disgrace for the amateur radio, however, I feel, if these people bring their crazy mumbo-jumbo, as usual, with denunciatory mails to institutions of the amateur radio and these – obviously from low motives – also still give place to it. This makes amateur radio look technically dumber than the majority of radio amateurs actually are.

Therein lies the real scandal.

[Receiver: Winradio Excalibur SIGMA, Antenna: MD-300DX-2x5m, Software: SDRC & Matlab]


Since scientific discussion on the website “The Search for MH370” under the half-megalomaniac subtitle “Serving the MH370 Global Community” is suppressed by its operator Richard Godfrey, a pensioner from Hesse, my blog is explicitly offered for corresponding discussions and rebuttals. It may be easier to successfully suppress facts in this matter with denunciatory e-mails to the DARC chairman – to write such mails is as much a question of character as to follow them at all – but nevertheless it can be exciting to learn something more about “alternative physics”.

Here is my comment suppressed on Godfrey’s website:

Hi – the new paper by Westphal et al. attempts to demonstrate a rocket start from WSPR log data. However, it is not clear from the paper in which way this proof should have succeeded. Apparently, “SNR anomalies” of the HF propagation are used for this purpose without distinguishing the term “anomalies” from the “normal case”. It would be interesting if the authors could explain exactly this in a comprehensible way. I am also asked for the long announced paper, in which the authors wanted to dedicate themselves to HF propagation and aircraft scatter and in which hopefully also representations/calculations of the radar cross section find entrance. There is a lot of preliminary work on this that meets scientific standards – and the community is now eagerly awaiting a methodology that will make this surprisingly possible for WSPR log data as well.
How strong the dynamics of the ionosphere are, and that Aircraft Scatter is clearly not detectable in the sum signal, but only by FFT analysis, I have presented in my latest blog entry:
73 Nils, DK8OK


  • Kudos to you Nils, for taking up the fight against self-proclaimed experts trying to get their 15 minutes of fame based on preposterous claims that don’t hold water to anyone even only superficially interested or educated in the physics of propagation (like me).

    • … thanks, Ollie – and it is not so much the “experts” whom I would like to reach with it – knowing that for this not methods of physics but those of psychology would be necessary – but those who eagerly parrot it. I had expected that at least the community of the radio amateurs, imagining themselves always at the spearhead of the technical-scientific development, would counter this Godfrey’s nonsense. But unfortunately the absolute opposite is the case: Fools support fools. Of which later, on occasion.
      73 Nils, DK8OK

      • Well done Nils. Godfrey is an attention seeking crackpot. I have been a Ham for over 40 years in G and VK. I saw Godfrey on TV with this nonsense and immediately dismissed his claims as impossible. A basic understanding of HF propagation with its myriad of natural atmospheric effects and the limited information in the WSPR database was enough for me to dismiss the idea without the need for further examination. Any Ham who falls for this pseudo-science must have the IQ of a pencil. I am amazed that someone with Godfrey’s background can show such poor science. It makes me suspect a hidden motive.

      • … thanks, Geoff. And the biggest secret, indeed, seems why people of some background like Godfrey, Westphal (who recently strongly weakens and relativizes his theses) avoid to publish their shlock science in scientific papers for a discussion … 73 Nils, DK8OK

  • As a radio engineer and radio amateur, I applaud your work here to dispel this myth that the WSPR dataset allows one to determine the location of aircraft. As soon as I learned of this theory, I wanted to understand how this could possibly work (or not work). Having listened to many deep fades in my lifetime, you have confirmed my hunch.

    • … thanks, Vivian, and I am just preparing another entry on this topic which will be published in the last decade of April. The best of the publications of Godefrey et al. is that they sparked some interest in and work on propagation … 73 Nils, DK8OK

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s