Monday, 17 March 2014

Open Question: Could US Military SIGINT satellites help to narrow down flight MH370's last location?

Please note: this post contains discussions of a highly speculative nature

Over the past days, it has become clear that the lost Malaysian Airlines flight MH370 has flown on for some 7 hours after contact was lost at 17:20 UT (March 7 UT, local March 8). This information comes from radio "ping-backs" of the aircraft's ACARS system received by the Inmarsat 3-F1 satellite, a geostationary communications satellite that is located at longitude 64 E over the Indian Ocean. These ping-backs were received hours after the last radio contact with the pilots and hours after the transponder was shut off, and indicate that the aircraft was still powered and 'alive' hours after it disappeared. A well written story at the CNN website gives backgrounds on the receptions and the system.

Position and footprint of Inmarsat 3-F1
click image to enlarge

In this post, I will briefly summarize how Inmarsat 3-F1 detected the aircraft and determined a wide arc where the aircraft could have been at that time. I will then explore whether additional signal receipts by classified US Military Signals Intelligence (SIGINT) satellites might perhaps have been possible. If such additional receptions exist (an open question!) they would enable to further narrow down the location of the last ping-back.

That will largely be a theoretical exercise, as so far there has been no word that the US SIGINT satellite constellation did detect these ping-backs. This post therefore entails a clear element of speculation, and the central question remains an explicit open question.

Backgrounds: 'Marco Polo' between an aircraft and a satellite

Someone in the aircraft shut off the radar transponder beacon and the active ACARS messaging system near 17:20 UT. Yet this did not fully disable the ACARS system. The system kept answering periodic "pings" by the Inmarsat 3-F1 (1996-020A) satellite. These "pings", basically a kind of "Marco?" message,  are periodically sent out by the satellite and when received by the aircraft ACARS antenna, the aircraft pings back a brief "handshake" basically saying "Polo!". While such a handshake does not contain clear information about where the aircraft is when the active ACARS is disabled, it does contain the aircraft ID.

According to press reports, the last ping-back from flight MH370 was received 7 hours after the flight disappeared, near 00:11 UT on March 8. Apparently, only Inmarsat 3-F1 received these ping-backs.

From the time it took the radio-ping to travel from Inmarsat 3-F1 to the aircraft and then back again, the distance (but not direction) of the aircraft to the satellite can be determined. For example, at a radiowave speed of 300 000 km/s, a time difference of say 0.2 seconds between Inmarsat sending the ping and receiving the answer back, indicates the aircraft is at a distance of 30 000 km from the satellite.

Once you know the distance, you can draw a globe with that radius around the location of the Inmarsat satellite. Where that globe cuts the earth surface, it creates a circle, centred on the sub-satellite point. The aircraft must have been somewhere on that circle. This is basically how the wide arc that has been published was constructed, an arc which runs from Thailand to Kazakhstan in the north, and Indonesia to Australia and the Indian Ocean in the south. The aircraft could have been anywhere on that big arc, an area stretching thousands of kilometers.

To pinpoint the aircraft more accurately to a particular spot in the arc, one needs a detection by a second and preferably a third satellite.

Could US SIGINT satellites provide additional receptions for these pings?

One source of such additional ping-back signal receptions, in theory could be one of several Signals Intelligence (SIGINT) satellites employed by the US military. Please note that I say IN THEORY as the US government hasn't provided any statements that they did (which might indicate that they didn't). In other words: I am speculating on an open question here.

It depends on a lot of factors, not the least of which are questions whether these satellites were listening at the time, and whether they were monitoring the particular VHF/UHF radiofrequencies in question. Those are questions I do not have the answers to. What I will do, is discuss which US military satellites could potentially have received these ping-backs because they had coverage of the area.

1. The Mentor and Trumpet SIGINT satellites

Two US SIGINT systems in high orbits cover(ed) the relevant area: (1) several of the very large Mentor/Advanced Orion SIGINT satellites in geostationary orbit: and (2) one of the SBIRS/TRUMPET combined SIGINT and SBIRS satellites which moves in a Highly Elliptical Orbit and hovered high above the northern hemisphere at the time.

These SIGINT satellites serve to eavesdrop on radio communications including satellite- and mobile telephony, missile telemetry and signals from groundbased and airborne radar systems.

USA 184 TRUMPET imaged on 25 Aug 2009 by the author

 Mentor 4 imaged on 18 Nov 2012 by the author

The TRUMPET satellite in HEO which had coverage of (a part of) the area at that time is  USA 184 (2006-027A). The geostationary Mentor satellites covering the area are Mentor 1, 3, 4, 5 and 6 (1995-022A, 2003-041A, 2009-001A, 2010-063A and 2012-034A).

Position of various Mentor satellites and TRUMPET USA 184
Mentor satellite footprints

USA 184 area coverage and footprint detail
click image to enlarge

2. NOSS (Naval Ocean Surveillance System) SIGINT satellites

Apart from the Mentor and Trumpet SIGINT satellites in high orbits, the US also operates a series of SIGINT satellites with accurate geolocalization capabilities in a Low Earth Orbit. It concerns the US Navy Naval Ocean Surveillance System (NOSS) satellites, of which there are several. They operate in close pairs, orbiting at an altitude of about 1000 x 1200 km in 63 degree inclined orbits. Their main purpose is to locate and track shipping through the radio communications of the latter.

A NOSS duo (NOSS 3-4) imaged by the author on 29 Jan 2011

Two duo's of NOSS satellites were covering the northern half of the area at the time of the last ping-back received by Inmarsat 3-F1: the NOSS 3-5 and NOSS 3-6 duo's (2011-014A and B and 2012-048A and P).

The NOSS 3-6 duo had the best coverage, which includes the full northern arc from Thailand to Kazakhstan determined by the Inmarsat reception:

click images to enlarge
position of the NOSS 3-5 and NOSS 3-6 duo at the time of the last pingback

in 3D: yellow arc is where the aircraft could be according to the Inmarsat 3-F1 reception

Chinese SIGINT

China operates a satellite system similar to the US NOSS, consisting of three satellite trio's in the Yaogan series (Yaogan 9A, B, C; 16A, B, C; 17A, B, C). None of these however had coverage of the relevant areas in the Indian Ocean, central Asia or southern Eurasia at that time.

Coverage summary

From the brief satellite coverage analysis summed up above, it seems that the northern overland arc from Thailand to Kazakhstan was potentially well covered by various US military SIGINT satellites: five Mentor satellites, a TRUMPET and a NOSS duo. The southern Indian Ocean arc is slightly less well covered (no TRUMPET or NOSS coverage) but was nevertheless in view of several geostationary Mentor SIGINT satellites.

The question now is: could one or more of these SIGINT satellites have captured the same ACARS ping-backs received by Inmarsat 3-F1? If so, the combination of their data with the Inmarsat data could potentially narrow down the last known position of the aircraft considerably.

It all depends on whether the satellites in question were actively listening at that time, and moreover, whether their monitoring includes the radio frequencies in which the ACARS ping-backs of flight MH370 operated. It perhaps also includes questions like whether any signals received are all kept on file, or if some selection is made and much deemed of no interest is directly discarded.

Those are some big serious "ifs", that I simply do not know the answers to: this stuff is, after all, classified. So far, the US government has not indicated that one of their SIGINT systems did capture the ping-backs. Which might mean that they didn't, as I can't imagine that they did not check for it.

Classified SIGINT satellite positions in this post (and previous posts) are based on orbits calculated by Mike McCants, based on amateur observations communicated on the SeeSat-L mailing list.

Addendum 18 March 2014:
In my initial analysis posted 17/03/2014, I forgot to include two other and older geostationary US SIGINT satellites: the two Mercury/ADVANCED VORTEX satellites that are located over East Africa.

 click images to enlarge

It concerns Mercury 1  (1994-054A) and Mercury 2 (1996-026A). Both satellites were recently moved to a new orbital position over East Africa and are station-keeping there, indicating they are operational. Their footprint includes the area of interest, although the southern Indian Ocean arc is close to the edge of their coverage.

Mercury 1 imaged by the author on 29 Dec 2013


Thomas Tyldum said...

Very interesting, thanks!

But what I wonder about, is where are those other "arcs" from the 5 missing pings? It wouldn't reveal position of aircraft, but it would reveal if airplane was flying in a straight line or not the last five hours, and also reveal if last observations on radar can be trusted in such detail. I tried google this, but came no closer.

PapaGeorge said...

Katherine Harris, Fox News reported "....that's based on additional signal intelligence beyond that satellite link communications." Thought that might add grist to you speculation.

Aireps said...

Many thanks for your insights on the presence of SIGINT sats.

On 13 March, Malaysia's Transport Minister said that the last ACARS transmission was received from the aircraft at 17:07 UTC. ACARS was somehow disabled. Reportedly, the plane's SATCOM system kept on replying to Inmarsat's SATCOM pings for several hours. Therefore, it would be better to address SATCOM pings, not ACARS pings. You may also view info on ACARS and SATCOM pings

Tony Verow, MD said...

Great explanations, thank you for the hard work. I suspect that there may be issues in gleaning data from the fact that the southern Indian Ocean is not terribly relevant most of the time to US military and geopolitical concerns. I would hypothesize that American satellite resources are usually deliberately spent on the Middle East, North America, etc. Even if signals are picked up from the Indian ocean they may not be high on the list of priority in terms of processing capacity, analyst time, etc. Nonetheless there might be very useful information stored somewhere !