A new paper on the search for MH370 and WSPR technology based on the public quotes from
Brian Anderson
Ed Anderson
John Chambers
Mike Eichhorn
Mike Exner
George Gatehouse
Richard Godfrey
Frank HolzƤpfel
Victor Iannello
Ari Joki et al.
Arto Lehtolainen
Jianbing Li et al.
Jean Luc Marchand
John Moore
Duncan Steel
Niels Tas
Bill Tracy
Robert Westphal
Dennis Workman (deceased)
and the MH370 Families.
The paper can be downloaded here
@All
A new article by Geoffrey Thomas examines the pro and contra of using WSPR technology in the search for MH370:
https://www.airlineratings.com/news/mh370-new-report-electrifies-science-community-gives-hope-relatives/
What signals from what receivers where were recorded on 8 mar 2014 that may have something to do with MH370?
Nothing teceived after that date is relevant.
@All
A second article by Geoffrey Thomas looks at the key contributions in the search for MH370 over the last 7 years:
https://www.airlineratings.com/news/new-tracking-technology-report-mh370-deserves-support/
@All
As mentioned in my recent paper, it is the first in a series of three.
In the 1st paper I covered the final major turns of MH370 from the last civilian radar point at 18:00:51 UTC until the 4th Arc.
There were a number of turns and the MH370 hypothetical flight path followed did not match the Malaysian military radar trace presented at the Metro Lido Hotel, Beijing to the Chinese next of kin on 21st March 2014. The MH370 hypothetical flight path followed does match the Inmarsat satellite data between the 1st Arc at 18:28:15 UTC and the 4th Arc at 21:41:27 UTC.
Below is an example at 21:42 UTC just after the 4th Arc in a regional view.
https://www.dropbox.com/s/cjwdltk0hg8brse/GDTAAA%20WSPR%20Links%20MH370%202142%20UTC%20Regional.png?dl=0
The next and 2nd paper will cover the initial flight of MH370 from take off until loss of civilian radar contact at 18:00:51 UTC.
This will cover the diversion and up to waypoint BITOD.
Below is an example at 17:20 UTC just before the diversion in a local view.
https://www.dropbox.com/s/atdyxlg0r2uaqs8/GDTAAA%20WSPR%20Links%20MH370%201720%20UTC%20Local.png?dl=0
The 3rd and final paper will cover the end of flight of MH370 from after the FMT turns are completed at the 4th Arc up until the 7th Arc and perhaps a few minutes beyond.
This will cover the crash location.
Below is an example at 00:22 UTC just after the 7th Arc in a local view.
https://www.dropbox.com/s/su89wy85hailfne/GDTAAA%20WSPR%20Links%20MH370%200022%20UTC%20Local.png?dl=0
Below is an example at 00:22 UTC just after the 7th Arc in a global view.
https://www.dropbox.com/s/ehj0v0sycagtkn7/GDTAAA%20WSPR%20Links%20MH370%200022%20UTC%20Global.png?dl=0
Disclaimer: The final major turns of MH370 are published as a paper and characterised as a working hypothesis. The initial flight, GDTAAA validation and end of flight of MH3700 are not yet published and any initial results are characterised as speculation.
@All
The physics behind GDTAAA and this latest WSPR technology is based on both the aircraft forward scatter doppler as described by the Ari Joki et al. in their paper as well as the aircraft vortex wave behind and to the sides of the aircraft as described by Jianbing Li et al. and Frank Hƶlzapfel in their papers.
It is possible to see both effects in the detection of flight SIA352 at 16:40 UTC:
https://www.dropbox.com/s/6bcl4vwq1hmuca4/GDTAAA%20WSPR%20Links%20MH370%201640%20UTC%20Local.png?dl=0
Two minutes later at 16:42 UTC the aircraft has moved on and there are no detections:
https://www.dropbox.com/s/r2ul6fdulwavbev/GDTAAA%20WSPR%20Links%20MH370%201642%20UTC%20Local.png?dl=0
At 16:44 UTC there is a possible detection of the wake to the side of flight ETD474A:
https://www.dropbox.com/s/ow3gj0nl08zjgcs/GDTAAA%20WSPR%20Links%20MH370%201644%20UTC%20Local.png?dl=0
At 16:46 UTC flight ETD474A is detected again as well as flight UAL895:
https://www.dropbox.com/s/fainn6xe54xt0nu/GDTAAA%20WSPR%20Links%20MH370%201646%20UTC%20Local.png?dl=0
At 16:48 UTC flight ETD474A is detected again and multiple detections now allow a tracking of the aircraft. In addition flight SIA322 and flight MAS052 are detected:
https://www.dropbox.com/s/qu1r4h3ugxa5jh4/GDTAAA%20WSPR%20Links%20MH370%201648%20UTC%20Local.png?dl=0
At 16:50 UTC flight ETD474A is further tracked and flight UAL895 is detected again and can start to be tracked as well:
https://www.dropbox.com/s/9kfwu1klxk5v4v6/GDTAAA%20WSPR%20Links%20MH370%201650%20UTC%20Local.png?dl=0
Disclaimer: The final major turns of MH370 are published as a paper and characterised as a working hypothesis. The initial flight, GDTAAA validation and end of flight of MH3700 are not yet published and any initial results are characterised as speculation.
The paper by Ari Joki et al. describes the results of experiments using a 500,000 watt broadcast transmitter 4300 km from the receiver (direct path). That station also uses a high gain antenna pointed to Europe. Thus, the EIRP is ~70 dB greater (10,000,000 X the power) than the typical WSPR transmitter. All the aircraft Joki detected, by observing Doppler shift on a waterfall graph, were near the receiver, within ~200km (LOS). Under those conditions, an aircraft can be detected (but not tracked without multiple baselines).
But WSPR is nothing like that. It does not record the data necessary to produce the waterfall (raw I-Q samples), and MH370 was nowhere near the TX or RCV antennas, a fundamental requirement. That makes it impossible to detect, much less track MH370 using WSPR observations.
@ALSM
You are missing the point.
Ari Joki et al. state: “For air traffic safety augmentation one possibility is to take advantage of existing radio amateur and dx- listener resources. During the periods where the proprietor is not using the receiving equipment for their hobby, perhaps the receiver could be utilized as part of an air traffic monitoring network.”
Why do you think they recommend to take advantage of existing radio amateur and dx listener resources as part of an air traffic monitoring network, if as you falsely claim it is impossible to detect or track aircraft using existing radio amateur resources?
Ari Joki et al. explain that air traffic monitoring over sparsely populated or ocean areas has similar challenges to homeland security long-distance early warning. In both cases the need is to detect and track relatively far away airborne objects. High measurement accuracy is not very important. Poor detections are better than none.
The more conventional microwave radar bands suffer from radio horizon limitations. The frequencies where propagation follows earthās curvature are congested by many users. No one is suggesting this is easy.
It is an obvious idea to see if it is possible to take advantage of the radio amateur community to detect and track aircraft including MH370.
Dear Mr Richard,
Title: Attaching Underwater Ground Penetrating Radar To Towed Side Scan Sonar of Mh370 Search
I really shall say this will be first paper publication of Underwater Ground Penetrating Radar in lowest depth of several kilometres underwater, this is a breakthough to scientific community and also, by attracting those researchers to Mh370 search, researchers of ground penetrating radar, with the use of attachment of underwater ground penetrating radar to towed side scan sonar, please ask researcher to use 2 to 5 km range underwater ground penetrating radar, make good underwater casing to it and use longer cable for the towed side scan sonar so it can be lowered to the range of underwater ground penetrating radar. Just like Mr Gibson said, they missed the search of the particular coordinates because of existance of canyon and trench structure, by using the attachment of underwater ground penetrating radar and towed side scan sonar the problem can be solved. Click links for reference:
https://www.malagpr.com.au/underwater-gpr.html
https://www.sciencedirect.com/science/article/abs/pii/S0926985117305268
Because the attachment of underwater ground penetrating radar to towed side scan sonar is new, use trial and error method, method of underwater ground penetrating radar switched on with towed side scan sonar switched off and method of underwater ground penetrating radar switched on with towed side scan sonar switched on. So basically the towed side scan sonar only aid in delivering the underwater ground penetrating radar to its range and supply electrity if needed.
Best regards,
Lai Shu Ying
@Lai Shu Ying
Welcome to the blog!
It is speculation that MH370 may have been missed due to the difficult sea floor terrain including volcanos, sea mounts, trenches, canyons and crevices.
I expect that large items such as engine cores and landing gear will still be detectable on the sea floor after 7 years.
Your Ground Penetrating Radar may be very useful if adapted to the large depths involved especially when the sea floor is covered with a thick layer of silt or mudslides have covered debris over time.
Richard, the WSPR/GDTAAA hypothesis is the smartest, most exciting and innovative thing I’ve seen in the past couple of years. The concept is one of brilliant simplicity. You and co-developers of this concept deserve all the kudos and wider (research-)community support to work out this concept.
RF propagation is a subtle and often counter-intuitive thing. It is a plain fact that the radio-photons will interact with & propagate around the curvature of the Earth in the ionosphere. By the sheer number of photons even in the low-power applications of WSPR radios these photons will interact with anything and bounce of anything on & off the Earth. So please don’t let the nay-sayers get you down when they throw one-liners or freshman physics at you that completely fails to capture the complexity of the problem.
HAM people are enthusiastic and sympathetic folk, but most of them aren’t physicists and are only able to visualize electromagnetic fields as inane simplifications. Of course they don’t want to hear that…
So you are right to put the detailed physics aside and try to see what you can distill from an _information_ point of view, see if some plausible correlations can be made with any degree of certainty.
However, that path – bold as it is – is also fraught with peril.
One piece of very relevant physics that is ignored is your assumption of great circle connections between tx/rx points. I’m not arguing it is invalid (I don’t know), but your argument that this assumption holds would be much stronger if you can enhance your analysis to exclude multi-path reasons for the disturbances to which you attach the significance (“tripwire”).
By sticking to this assumption, you’re open to hard to disprove claims of cherry-picking. Correlation isn’t necessarily causation, as you know. It might be easy to demonstrate that you’re mostly just fitting noise onto external constraints (i.e, INMARSAT data, or ADS-B in case of the validation papers).
And this counterargument is very hard to shake off if you don’t provide full disclosure of all the relevant data, methods, and code.
More worrying, from my layman’s vantage point:
You conclude on one hand a significant deviation from established flight path, yet at the same time this deviated path is the one that also happens to end up in the same place.
Casual lurkers like me know that this established path was based on minimum energy constraints (max. fuel saving), so that seems to be in flagrant contradiction to your path.
Next, much more worrying, the path isn’t lining up with the official millitary radar data. At one point in the dropbox comments you simply discard that radar data as “faked” without any evidence to back this up.
I’m afraid WSPR analysis won’t be taken seriously if you make such unsupported bold claims. Just because the raw data hasn’t been released yet doesn’t make it faked. Even raw data can be faked. This is a can of worms you shouldn’t be getting into…
Best wishes and good luck with the analysis, which I will follow with fascinated interest!
And more
@Hendrik,
Welcome to the blog!
Many thanks for your kind words, your warnings and your worries. It sounds like www stands for words.warnings.worries not world.wide.web. š
Using WSPRnet data to detect and track aircraft is a working hypothesis in its infancy. GDTAAA as a software tool is still in development and a number of tests still need to be performed. It will take time to fully test the system and the process. The limitations of GDTAAA are not yet fully understood and the software tool is far from being characterised as a mature product.
We have written 4 papers on the subject since March 2021:
1. Geocaching in the Ionosphere.
2. Using the WSPRnet and Inmarsat Satellite data in the Search for MH370.
3. Global Detection and Tracking of Aircraft as used in the Search for MH370.
4. When Do Whispers Shout.
The physics behind GDTAAA and this latest WSPR technology is based on both the aircraft forward scatter doppler as described by the Ari Joki et al. in their paper as well as the aircraft vortex wave behind and to the sides of the aircraft as described by Jianbing Li et al. and Frank Hƶlzapfel in their papers. I have not yet completed a literature review.
I note that you quote the highly critical VK2IJM who has previously stated on this blog ācorrelation isnāt necessarily causationā. I have made it clear that WSPR signal anomalies may be for a variety of reasons and only one of those is related to an aircraft. It is not yet clear whether the one possible relation between WSPR anomalies and aircraft detection is directly due to the physical aircraft object or the forward scatter or the wake or all or some or none of the above. There may be a correlation. There may be an indirect causation. There may be a direct causation. There may be a correlation without causation. There may be a causation without a correlation.
I am not assuming that an individual position or progress indicator has any meaning. A series of indicators that align with the predicted flight path of an aircraft over time may be significant. I am not cherry picking, I am predicting where the aircraft will be two minutes further on, for a given ground speed and given track. I then observe if there are any WSPR indicators. I am doing validation tests to find out what the capabilities and limitations of GDTAAA are. My patent application for GDTAAA has been accepted, but it takes time to gain approval.
You demand my method, data and code. You appear to have missed the 4 papers and 140 comments on this blog discussing method, data and code. For example, I have published on this blog 5 Excel tables with all my working WSPRnet derived data for each scenario I have looked at, including the data enrichment and the code generation. You also appear to have missed the plan for future enhancements including propagation analysis and ray tracing. Your criticisms are unfounded.
Here is a recent example taken from the AMSA Challenger SAR flight VH-XNF to D2S2 on 14th May 2021 at 04:00 UTC:
1. WSPRnet data enriched and filtered together with code generation (already published):
https://www.dropbox.com/s/4b02f8rgipdnha9/wsprspots-2021-05-14-0400-0430%20UTC%20Anomalous.xlsx?dl=0
2. MATLAB shell with code generation (new publication):
https://www.dropbox.com/s/8p8jk2yt1apvpep/GDTAAA%20Sample%20Code.pdf?dl=0
3. MATLAB function call generating Great Circle output on WGS84 Oblate Spheroid (new publication):
https://www.dropbox.com/s/8fpkwe5mmn2f3cv/GDTAAA%20Function%20Call.pdf?dl=0
4. GDTAAA output (already published):
https://www.dropbox.com/s/r5a531ju4fk11r9/GDTAAA%20AMSA%20VH-XNF%2014MAY2021%200400%20UTC%20A.png?dl=0
You claim that I provided no evidence of my flight path of MH370 in the Malacca Strait. I wrote a whole paper on the subject from the last civilian radar contact to the 4th Arc. In that paper, I have announced two further papers: one paper on the initial flight of MH370 up to the last civilian radar contact and a second paper on the final flight path from the 4th Arc to the 7th Arc and beyond to the crash location. These papers will only be published when the validation testing is complete. Any advance public announcements by you of what you think I will write are premature. Unfortunately, you are not the only one speaking on my behalf without authorisation. Please do not eaves drop on conversations, where other analysts ask me questions on my Dropbox site rather than by email or on my blog. Please just pretend my Dropbox is private.
This blog is not a court of law and I refuse to be dragged into discussions of a legal nature.
Thank you for wishing me luck, even if half heartedly, soaked in warnings and worries! š
Why not join the research team? š¤Ŗš¤Øš§
Dear Richard,
Thanks for the very detailed follow-up!
The correlation/causation is just a well known maxim to caution against too much enthusiasm or tunnel vision š I haven’t actually read that person’s comment whom you refer to.
I’m fully with you on your approach, and I think this makes perfect sense. I can see that the physics behind it can be a completely parallel tract.
Funnily it was Oliver Heaviside who once remarked: “why should I refuse a good lunch if I don’t fully understand the details of the intestinal tract?” in response to his mathematical detractors on operational calculus (fully correct, later shown to be equivalent to Laplace’s transform).
He is also the one who co-postulated the existence of the ionosphere and the one who reformulated and simplified “Maxwell”‘s equations of electromagnetic field theory in the form thought in universities today, from the incomprehensible 22-variable quaternion potentials of Maxwell.
I’m only saying: this path you take, while both bold and promising, is also prone to substantial criticism, so full and complete transparency are necessary to gain conviction. For example, I quickly note that the matplots you posted are just generating plotting output, it’s not an analysis. And I understand your considerations / patent applications – difficult balance.
While I appreciate your call to join the team I don’t consider myself a competent enough statistician :-), I just wanted to point out the obvious as an outsider, something that might be useful to take into account.
Also, my apologies for dropping in on what you considered to be a breach of privacy, I didn’t realize you considered them private.
After writing my first comment I realized my first point – may actually be OK – the two different flight paths might not be incompatible – the original hypothesis is for a late turn south, much further to the north-east. So that might balance out the more curved path you found.
Regarding the radar in-or-consistency, I’m sure you have your reasons and we’ll learn in due time.
My main reason for writing is to express my admiration & happiness for reading about the new method and analysis and to sincerely wish you best luck, and hope the community will take it up.
Cheers,
Hendrik
I’m writing to say
@Hendrik Beijeman,
Many thanks for your comment!
I am very aware that the path I have take is prone to substantial criticism. There are some nay sayers who have been vocal in a disinformation campaign on nearly all the MH370 community’s web sites. They come across to me like members of the Flat Earth Society. In Covid-19 times I have developed a thick Pangolin skin.
My local pub is now debating correlation and causation! I am told the answer is 42. Another guy promises me to stay up all night and deliver the answer to WSPR and aircraft detection by midday tomorrow. We will see.
I agree with Oliver Heaviside. “Why should I refuse a good lunch if I donāt fully understand the details of the intestinal tract?ā Although at odds with the scientific establishment for most of his life, Heaviside changed the face of telecommunications, mathematics, and science (according to Wikipedia). Not bad!
Patent aside, I am only looking at SNR and Drift anomalies in the WSPR data. I filter out WSPR links within the same 2 character Maidenhead Grid Cell because they are very likely short path links and not likely to be crossing the southern Indian Ocean (which is my MH370 are of interest). I also filter out WSPR links less than a total of 12 characters on the Maidenhead Grid as they will not be accurate. I do employ an algorithm based on the short path distance to differentiate between SNR anomalies over shorter or longer distances. I have not yet found the need to develop more complex algorithms, although that was my expectation at the outset. I expect those will come.
Sorry we cannot win you to the team. My own statistics does not go much beyond “Statistics” 4th Edition by David Freedman.
@Richard,
Assume the WSPR method is valid. Then the resulting flight path must still meet the constraints of the inmarsat arcs and flight models (i.e, fuel consumption, flight envelope, etc). The original DST Bayesian analysis assumed (?) a final flight path with constant heading / altitude which led to the probability map with a peak near 38S/88E. What would be the range along the 7th arc if we let go all assumptions about flight path (so long they are physically plausible)? Are you aware of any such study, and if so, is it compatible with your findings about the WSPR-derived flight path?
Cheers,
Hendrik
@Hendrik,
The WSPR method is a work in progress and is being continually refined.
In my first paper on WSPR, I made it clear that it is the combination of WSPR with the Inmarsat satellite data, 9M-MRO fuel data and Boeing 777-200ER performance data that will give more precise information of the MH370 flight path and crash location.
Bayesian Methods in the Search for MH370 assumes an early FMT, high and fast flight path and infinite fuel.
Many analysts have considered flight paths unconstrained by the false assumption that the flight path has to be straight and level. Flight paths based on waypoints, magnetic headings, single engine operation and other criteria have been proposed. I am not aware of any flight path that is compatible with the WSPR-derived flight path I have presented.
@Paul Dent
Welcome to the blog!
You ask the key question: “What signals from what receivers were recorded on 8 mar 2014 that may have something to do with MH370?”
The WSPRnet database records 91,895 signals during the flight of MH370 between 7th March 2014 16:42 UTC and 8th March 2014 00:22 UTC.
Out of these, I have selected 58,696 signals that may be relevant:
1. Distance between transmitter and receiver ā„ 1,000 km. Distances recorded are between 1,000 km and 19,437 km.
2. Transmitter power ā„ 10 dBm.
3. Transmitter frequency band between 3 MHz and 30 MHz.
For the maximum precision of the propagation path between transmitter and receiver, I only select those WSPRnet links where both the transmitter and receiver Maidenhead grid code is 6 characters in length, leaving 51,487 WSPRnet links.
There are 9,710 (19%) WSPRnet links that show a signal level anomaly.
There are 13,202 (26%) WSPRnet links that show a frequency drift anomaly.
There are 2,412 (5%) WSPRnet links that show both a signal level and frequency drift anomaly.
There are 304 different WSPRnet receivers involved and 7,533 unique transmitter-receiver link pairs.
An Excel table with a list of all the 51,847 WSPRnet links, sorted by receiver, is attached (download 5.6 MB):
https://www.dropbox.com/s/znrmry80h4e8xnz/WSPR%20Cells%2007032014%201642%20UTC%2008032014%200022%20UTC%20MH370%20Receivers.xlsx?dl=0
I agree with you that signals at other times are not relevant to MH370, but they are relevant to detecting and tracking other aircraft in the Indian Ocean region both in 2014 and at other times. I am using other flights to test the hypothesis that WSPRnet data can be used to detect and track MH370.
Great research. There has been such a huge leap in signal processing technology currently used in astrophysics, radio astronomy, and microbiology.
I don’t know enough (anything?) about how to do this, but these two wrote a lot of processing software for the European LOFAR project: Frits Sweijen & Alexandar Mechev. I wonder if they could look at the WSPR data from a different perspective?
@Roy Dugger,
Welcome to the blog!
I agree there has been a huge leap in digital signal processing technology.
Feel free to put Frits Sweijen & Alexandar Mechev in contact with me.
richard@mh370search.com