Searching for MH370
Aircraft Scatter Data Not Conclusive
Aircraft-Scatter, WSPR and MH370: A Critical Appraisal
Time and again, there are news stories in the professional and popular press about the fact that log data from the WSPR data network can help locate aircraft. In particular, an effort has been undertaken to determine the actual crash site of Malaysian Airlines Flight MH370. This effort essentially amounts to detecting ‘unusual’ level-jumps and frequency changes (‘drift’) in the archived WSPR log data and attributing them to reflections from specific aircraft (‘aircraft scatter’).
In his latest blog entry, RadioUser contributor Nils Schiffhauer, DK8OK, for the first time evaluates this theory more critically. His findings are based on years of observation of aircraft scatter on short wave, as well as on an investigation of about 30 Doppler tracks. The results of this complex analysis of more than 10,000 data in one example alone are sobering: The effects of aircraft scatter on the overall signal are almost always well below 0.3 dB.
To prove a correlation between level changes of the overall signal and aircraft scatter seems hardly possible based on the WSPR data material. The reasons are manifold but lie mainly in short wave propagation, where level changes of 30 dB within a few seconds are the rule rather than the exception.
However, since the local and temporal state of the ionosphere is not known in previous investigations on the WSPR data material - it is recorded in parallel in professional OTH radar systems and calculated out of the received signal - level jumps cannot be unambiguously assigned from the sum signal alone.
This finding is supported by further arguments offered in Nils’s blog (at the URL below). The illustrations show the carrier of China Radio International (CRI), which shows – under a 100Hz FFT – magnifier some aircraft Doppler traces, and the separation of Carrier, Doppler and Noise (at a bandwidth of 1Hz) to analyse each effect on its own, plus any dependency.
(Source: Nils Schiffhauer)