WSPRnet radio signals can reliably detect and track aircraft over long distances to the other side of the globe. Anomalies in the WSPRnet data, in either the received signal level, or received frequency, or frequency drift indicate a possible disturbance by an aircraft.
WSPRnet is a multi-static and multi-frequency system with global coverage. There are currently around 6 million distinctive links between WSPR transmitters and receivers from around the world with a propagation distance greater than 3,000 km recorded in the WSPRnet database.
The long coherent integration time of the WSPRnet receivers, the enhanced radar footprint of modern aircraft in the WSPRnet wavelength bands and the global coverage of WSPRnet propagations ensure a high level of detection and reliable tracking of aircraft.
The WSPRnet technique is, in its simplest form, statistical post processing of the meta-data from a communications link. We combine the reflection of radio waves by aircraft as used in radar, with ionospheric propagation and the WSPR protocol to detect and track aircraft over long distances.
This technique was used to track MH370 from the last confirmed radar detection on 7th March 2014 at 18:00 UTC until the end of flight on 8th March 2014 at 00:28 UTC. A total of 313 anomalies in the signal level or frequency of the WSPR signals were detected at the estimated position of MH370 at 130 different points in time.
The crash location of MH370 was at 29.128°S 99.934°E, which is outside the previous ATSB and Ocean Infinity underwater search areas.
A technical paper describing the WSPR technique for detecting and tracking aircraft over long distances can be downloaded here
@Kurt Kastenmeyer,
You write: “I am aware there are a number of solutions that match the satellite data for the final leg into the Southern Indian Ocean. My understanding of the UGIB approach is that it employed a novel statistical analysis of uncorrelated parameters in addition to the basic satellite data. That analysis was paired with other data, including a fuel model, weather data and debris drift. The final analysis showed a much higher probability for the UGIB route over other routes that were also considered.”
You ask: “A couple of questions, if I may:
1. Did the other analysts you mentioned ever consider a variety of data from different sources, as used in the UGIB paper, or did they only consider the basic BTO/BFO data?
2. Has the UGIB approach been debunked? If so, where?”
I wrote the paper “The Final Resting Place of MH370” together with inputs collated from Bobby Ulich (193 submissions), Victor Iannello (65 submissions) and Andrew Banks (14 submissions). The paper was checked by all authors before publication. The paper was published on 9th March 2020 and has since become known as the UGIB paper, which stands for Ulich, Godfrey, Iannello and Banks. We agreed that Bobby Ulich would be the lead author, who had the final say should all authors not agree on any particular point.
The paper took 4 months to write and has 396 backing files. The paper was the result of 8,224 email exchanges between the various authors.
The paper is based on an analysis of 1,372 possible flight paths published on 24th June 2019 in a paper that I wrote as lead author together with Bobby Ulich and Victor Iannello titled “Blowin’ in the Wind – Scanning the Southern Indian Ocean for MH370”. The paper determined a possible flight path for MH370 ending near the 7th arc at around 34.4 °S.
The final analysis does not show “a much higher probability for the UGIB route over other routes that were also considered”, as you claim.
The UGIB paper uses three MH370 flight models independently developed by myself, Victor Iannello and Bobby Ulich, which we cross checked and ensured aligned with each other. We considered the Inmarsat paper titled “The Search for MH370” by Ashton et al. published in the Journal of Navigation and accepted on 14th September 2014 with an end point at 34.7°S 93.0°E. We also considered the DSTG paper titled “Bayesian Methods in the Search for MH370” published on 30th November 2015 with a search area between 35.5°S 92.0°E and 39.5°S 86.0°E centred on an end point at 38.0°S 88.0°E. Both these papers are acknowledged in the UGIB paper.
The UGIB paper did not include a comparison of all other MH370 analysts’ flight paths or end points. There are over 120 books published on MH370, each with their own theory from being abducted by aliens to being diverted to Kazakstan by the Russians or Diego Garcia by the Americans or shot down over the South China Sea by the Americans. The serious MH370 analysts variously use data from Malaysian Airlines, ATSB, Boeing, Inmarsat and Oceanographers, the leaked Royal Malaysian Police report, the leaked Civilian Primary Radar raw returns from KLIA TAR, Genting RSR, KB TAR, Butterworth TAR and Langkawi TAR, the official ACARS data and the leaked official ADS-B data from Terengganu Airport.
The UGIB paper included two novel aspects, firstly an accurate fuel model based on MAS Engineering data from previous flights by the actual aircraft registration 9M-MRO and secondly a statistical method using up to eleven BTO, BFO, Along-Track and Cross-Track Position Error statistics recognising that the errors are uncorrelated with themselves and with each other (please see page 153 of UGIB). The paper also included an accurate weather model of the Southern Indian Ocean independently developed by myself, Victor Iannello and Bobby Ulich, which we cross checked and ensured aligned with each other.
UGIB also included a drift analysis as you correctly mention, but I would point out that none of the authors claim to be Oceanographers and the various MH370 drift analyses by major Oceanographic institutes do not agree on a MH370 end point.
In common with many other analysts, the UGIB paper has a major flaw, that the assumed flight path was a straight line as described in section 5.1.3 Assumptions: “In order to solve this problem using a single, predetermined route, it must be assumed that there were no lateral manoeuvres between 19:41:03 UTC and 00:11:00 UTC. The BTOs and BFOs vary smoothly with time during this interval, indicating the likelihood of no significant turns, speed or altitude changes.” This assumption is based on five Inmarsat satellite BTO observations initially at hourly intervals, where the last interval is closer to 90 minutes. An assumption based on so few observations made at large time intervals and over a large time span is inherently unreliable.
The issue of “straight line thinking” has been discussed at length in over 90 comments on this website. The first comment on straight line thinking was on 18th November 2021 at 21:01: “ATSB, DSTG, Inmarsat, Boeing, IG and just about every other analyst assumed that MH370 flew in a straight line due south.” This was followed by another comment and graphic:”Boeing, Inmarsat, ATSB, DSTG, Mick Gilbert, Captio, Ed Anderson, Victor Iannello and UGIB have all presented analyses of MH370 of the flight of MH370 following straight lines:”
https://www.dropbox.com/s/kpsh92ihs6xghwj/Straight%20Line%20or%20Not%3F.pdf?dl=0
A more recent comment on 13th March 2024 can be found at the following link:
https://www.mh370search.com/2023/02/26/the-ongoing-search-for-mh370/comment-page-2/#comment-2531
Another recent comment on 9th March 2024 can be found at the following link:
https://www.mh370search.com/2024/02/15/wspr-aircraft-tracking/comment-page-1/#comment-2484
In a guest post by CAPTIO titled “How many straight-line trajectories for MH370?”, published 20th April 2021 (after the UGIB paper was published), they analysed millions of possible straight line paths. They analysed millions of numerical estimations of such trajectories, which have been computed based on Arc2 latitude, track direction and speed limits using discrete sets of crossing locations at each arc. Further constraining the selection of trajectories by their ground speed standard deviation within the operationally acceptable limits, the number of acceptable linear trajectories is reduced to around 38,000.
You can read the guest post and download their paper at the following link:
https://www.mh370search.com/2021/04/20/guest-paper-by-captio/
If there are 38,000 realistic straight line paths, how can UGIB justify the assumption that they have published the only feasible straight line path?
If there are 38,000 realistic straight line paths, there are substantially more realistic flight paths that allow turns or speed changes or step climbs. How can UGIB justify that there were “no significant turns, speed or altitude changes”.
The invalid and unjustified assumption of a straight line flight path, invalidates the UGIB defined end point for MH370.
Dear Richard,
Congratulations on the impressive work on aircraft tracking using the WSPR network. I work in a different academic field, so the technicalities of this are lost to me, but I have learned a great deal from your website and papers, nonetheless.
Chapter 5 in the paper titled “How does WSPR detect Aircraft over long Distances?” displays the highly impressive reconstruction of a flight path using WSPR data. This has been done on an airplane that took off and landed safely, flight THY161. In the case of MH370, the community is looking for an airplane that did not land safely, moreover, is lost in the ocean.
My question to you:
Have you considered testing the WSPR data on Air France Flight 447 (AF447/AFR447) which on the first of June 2009 crashed into the Atlantic Ocean? The plane has been recovered and the crash location is known.
Kind regards,
Petra
Can the WSPR / GDTAA software detect clear air turbulence, ie does clear air turbulence create detectable anomalies in signal transmission / reception patterns that could be used to relay weather warnings to pilots in real time?
“Scientists say that clear air turbulence, which is invisible to radar, is getting worse because of the climate crisis.”
https://www.theguardian.com/world/article/2024/jul/01/air-europa-plane-turbulence-injuries
Any practical application of the WSPR / GDTAA software would help to raise its profile worldwide, and this in time might help to persuade the relevant authorities to permit searches for MH370 in the areas where it is most likely that it crashed.
If it hasn’t happened already, it might be worth reaching out to the ITF (The International Transport-Workers Federation), as they were very helpful in locating the wreckage of the MV Derbyshire, after many years of corporate and governmental obfuscation.
Part of the problem for MV Derbyshire was the quality of weather forecasts that it relied on.
“The cause of the sinking was only established 20 years after the tragedy. The lack of wreckage or survivors initially impeded a formal investigation by the British government and Typhoon Orchid was established as the official cause of the sinking.
After several years of campaigns by unions and victims’ relatives under the “Derbyshire Families Association” for the official opening of an investigation into the causes of the accident, as well as two separate incidents of deck plating (frame 65) cracking in two of Derbyshire’s sister ships which unveiled flaws in design modifications, ITF funded a deep-water search mission to find the wreck, which was eventually found on 8 June 1994 off Okinawa, Japan and led to the reopening of the case.
Following two expeditions, the initial investigation outcome in 1998 was that the ship sank because the lid to a store hatch on the fore deck had been left unsecured, thus attributing the incident to crew negligence. After several years and conflicting theories as to what led to the sinking, the official cause was announced in 2000 and was traced in design flaws, noting that rough seas damaged the vessel’s deck fittings and hatch covers, subsequently leading to the chain locker, forecastle and then the cargo holds being flooded one by one, sinking the ship in less than two minutes.
Due to the use of a flushed deck, rather than the fitting of a more conventional forecastle, the mountainous waves caused by Typhoon Orchid were allowed to sheer off the covers of small ventilation pipes near the bow, which, over the next two days, caused seawater to enter into the forward ballast section of the ship and draw the bow slowly into the water.
As the bow dropped lower in the water, the sea started damaging the first hatch cover, which was not designed to withstand such conditions and flooded, opening the way for the others to flood too and eventually sinking the ship. In this final version, there was neither evidence that crew was responsible for leaving hatch covers open nor that the deck plating was flawed.”
https://safety4sea.com/cm-mv-derbyshire-remembering-largest-british-ship-ever-lost-at-sea/
@TommyL,
Detection of severe weather events is improving, but climate change has been shown, by studies like the one from Reading University mentioned in the Guardian article you linked, to be the cause of an increase of turbulence on long haul flights.
On 21st May 2024, SQ321 suffered a sudden drop of 6,000 feet over the Andaman Sea and had to divert and make an emergency landing in Bangkok with one person dead and several other passengers and crew injured and suffering skull, brain and spine injuries.
A week later, eight people were taken to hospital after turbulence during a Qatar Airways flight from Doha to Ireland over Turkey.
On 1st July 2024, an Air Europa flight from Madrid to Montevideo dropped 3,000 feet suddenly and was forced to make an emergency landing at a Brazilian airport due to “severe turbulence” over the Atlantic Ocean.
As you mention radar cannot detect clear air turbulence. ADS-B data transmitted by the aircraft can record the sudden change in altitude and rate of descent.
Radar with a pulse width of typically around 1µs cannot detect the change in aircraft altitude caused by turbulence. WSPR integrates the received signal over 110.6 seconds and can detect the effect of turbulence on the aircraft. WSPR cannot forecast areas of turbulence, it can only detect the effect of turbulence on aircraft.
Re:”WSPR cannot forecast areas of turbulence, it can only detect the effect of turbulence on aircraft.”
Thanks Richard,
It looks as though you are saying that WSPR can only identify clear air turbulence after an aircraft has encountered it, which would, of course, be of no use as a warning system.
From what I’ve read on this blog over time, my understanding of the way WSPR data is used for aircraft tracking is that anomalies in the timing between transmission and reception can arise when a radio signal is
1. reflected off the surfaces of an aircraft
2. disturbed by the heat and exhaust from the engines
3. disturbed by the turbulence caused by the aircraft as it moves through the sky
What I was wondering was whether clear air turbulence on its own would be capable of creating detectable radio signal anomalies.
While WSPR alone may not be sufficient, might WSPR analysis be able to play a part in the development of a real time turbulence locator mapping system (ie not so much as a means of ‘predicting’, more as a means of identifyiing, locating and tracking areas of unusually high turbulence, whether in clear air or not.
While on the subject of weather radar warning systems, there is a recent entry in the maintenance log of MH370 that appears to refer to the weather radar system.
It might be another piece of equipment which, once the wreckage is located and recovered, will help to make sense of what happened to the aircraft.
Below is an extract from item 35 in the recent technical log entries (Appendix 1.6A) in the MH370 safety report
“S/N 4880484
– 02 March 2014
Status mssg WXR sys
02 March 2014
MMSG: 34-44002.
WXR transceiver (right) has an internal fault.
MSG erased from maintenance page.
WXR (R) CB (P11/E16)
Nil existing faults.
Pls eml and report.”