Hacker News Reader

Comment by sopooneo | Hacker News Reader

Hearing about all the painstaking analysis of this flight made me wonder something: If you look deeply enough, how frequent are inexplicable situations? That is, if you took all the info from a normal flight, and cut it off at the halfway point and gave it to a group of enthusiastic investigators, would they find aspects that contradicted or truly did not make sense? My guess is they would.

This is not to say MH370 was without incident, only that apparent contradiction may be unavoidable if you look closely enough at anything.

View on HN

It's unlikely, since about 98% of the Earth has transponder coverage, so you at least have position and what the aircraft think's it's altitude is, with pretty good precision and sample rate. It's only in a few spots in the middle the ocean that this doesn't exist.

This was also relevant in that Airbus crash from maybe 15 years ago between South America and Africa.

This is not a problem on the main to/from North America routes because there are enough populated islands along the way (Iceland, Greenland, Hawaii, Guam, etc)

Not sure what exactly you mean by "transponder coverage", but if you mean primary or secondary surveillance radar, this is not true for most of the Atlantic or Pacific.

Radar (both primary and secondary, i.e. transponders) as well as ADS-B require line-of-sight, and oceanic air traffic control has to make due without that, which means operating either via ADS-C (i.e. planes self-reporting their position, autonomously determined via GPS or inertial navigation, via satellite communications) or effectively dead reckoning, augmented by occasional position reports via HF radio, which is called procedural control:

https://en.wikipedia.org/wiki/Procedural_control

Space-based ADS-B is now a thing. It is used for coverage of the north Atlantic.

https://www.navcanada.ca/en/air-traffic/space-based-ads-b.as...

Yes, but note how this press release is mostly explaining what this technology could eventually be used for. I'm not aware of any ACCs actually using it for oceanic control today or having made an announcement of planning to do so in the future.

It's a very exciting technology, sure, but I think it's many years of testing and flawless operation as a secondary data source away from replacing ADS-C on oceanic routes.

I think closer to 80-90% of all landmass has "active" surveillance, with another 40-50% of sea / oceans.

Most of the South Pacific / South Atlantic / Southern Ocean (Antarctic); a lot of the North Pacific / North Atlantic / Indian / Arctic outside of the major islands and direct routes has nearly no coverage.

I think even this number is high. 80-90% sounds about right for the mainland US and Europe, but definitely isn't true elsewhere. Canada probably has ~50% radar coverage, and oceans have essentially 0% once you're beyond line of sight from from the coast.

(And this is if you're a jet flying at high altitude. If you're down where smaller aircraft fly, I can point to areas on the US west coast that don't have radar or radio coverage. Heck, there's parts of the mainland US that don't even have NEXRAD weather radar coverage.)

> Heck, there's parts of the mainland US that don't even have NEXRAD weather radar coverage.

Is that a good comparison, though? It seems like NOAA funding for the NEXRAD sites is perpetually inadequate. The one in Portland occasionally has outages that last a while. I would have guessed that civilian air control radar is better funded than NEXRAD, and military radar probably so well funded that every square inch of CONUS is covered by at least one.

Domestic civilian and military radar is the same thing. It's (mostly) all managed by the FAA, and the military just gets a copy of the civilian radar feed as part of the Joint Surveillance System. [1]

The limiting factor is geography. Radar coverage in the US is excellent, but the mountains out west make universal coverage hard, and there's some areas that just aren't cost effective to cover.

At 40,000 MSL, yes, everything in the US is covered. Below 10,000 MSL though, there are definite gaps - especially in the mountainous west. It's just physics, the mountains block the signal, and filling in all the gaps is cost prohibitive. Instead, they just tell aircraft to fly higher if they want radar service.

As for outages... ATC radar goes out all the time. You just never hear about it because they either switch to an adjacent, (slower) en-route radar site as part of CENRAP [2], or fall back to procedural control (voice position reports on the radio). Redundancy is baked into the system at many levels, both technical and human.

[1]: https://en.wikipedia.org/wiki/Joint_Surveillance_System

[2]: https://aviation.stackexchange.com/questions/9464/what-is-ce...

That's fascinating information, thanks for sharing! Off to lose some time diving into wikipedia...

That number can't be right. 98% of the earth doesn't have primary radar coverage, let alone secondary (transponder) coverage.

ATC radar is line of sight. At an altitude of 40,000 feet, you're beyond the radio horizon and lose coverage after about 380 miles - and that's the theoretical best case.

Oceanic air traffic control is primarily a non-radar environment, using procedural separation. Even over land, most of Canada (outside major cities and the airways between them) doesn't have radar coverage. There's a map showing the gaps here: http://ilanreich.com/Public_Pics/Alaska/Section%20One.htm

(The military does have over-the-horizon radar, but that's mostly used for missile defense. OTH radar isn't used for air traffic control.)

The newer systems that have come online over the past decade or so (partially in response to 370) actually use the iridium satelites

These systems are using ADS-B and are only designed for operational/informational purposes so far, as far as I know.

They are not used for air traffic control or other safety-critical purposes.

And for anyone who doesn't know; ADS-B is simply the aircraft voluntarily reporting its own GPS coordinates.

So it's not going to help at all in a MH370 scenario.

> ADS-B is simply the aircraft voluntarily reporting its own GPS coordinates

True, but that is ultimately the case for all flights outside of primary radar coverage.

If it's using Iridum or (more likely) Inmarsat, it would technically be ADS-C. (There is also a parallel system that allows satellites to detect existing ADS-B signals, developed by Aireon - but this is relatively new and not used globally.)

These absolutely can be used for air-traffic control. ADS-B has better temporal resolution than radar, and ADS-C is at least as good as manual position reports over the radio (used for procedural control over oceanic/non-radar airspace).

However, it's worth noting that these aren't really "transponders" (in aviation, that usually refers to beacon used for Mode-A/C secondary radar). Also, these are all active systems, dependent on the aircraft's electrical bus and easily disabled. In the case of MH370 the transponder _was_ powered down for some reason.

The only system that the pilot can't turn off is primary radar, but again, coverage for this system is limited once you leave populated areas.

I believe GP was thinking about Aireon and similar systems, which are similar to ADS-C at a surface level, but solve a different problem:

ADS-C is a two-way communications protocol; ATC can request a certain update interval and I believe also boundary conditions requiring extra location transmissions. If that communication fails, both the ATC and pilot are informed about the fact, and they can reliably switch to other means of communications (e.g. voice position reports over HF).

> These absolutely can be used for air-traffic control.

ADS-B is (at least for now) mostly an augmentation of primary and secondary surveillance radar, as well as something to give other aircraft better situational awareness (via ADS-B in), as I understand it. I don't think either ATC nor pilots are currently relying on it for safety-critical decisions.

The big problem is that it's always possible for some plane's ADS-B transmitter to fail (maliciously or otherwise). On land, you have secondary or primary radar to fall back to.

This applies doubly so to third-party (i.e. non-ATC/non-AOC) satellite-based relaying services, since not only the ADS-B transmitter can fail in that scenario (and ADS-B was not originally designed to be received by satellites), but the relay service could fail as well, and the failure mode would be much less visible to both parties than in ADS-C.

Are you referring to AirFrance 447?

https://en.wikipedia.org/wiki/Air_France_Flight_447

Even more, I imagine the average flight would look extraordinary. You would always hit some rare event if you looked hard enough. Rare events are so common! What are the odds that this many people on the plane were military / government / communists / CEOs / Jews / elderly? Well, pretty high if you're looking after-the-fact. And if you look into the background of every person and started reaching out a bit it would get weirder. You'd have tons of former whatevers, brother-in-laws of important people, criminals, all sorts. And people would be absolutely certain it meant something.

In the case of MH370, this is why I always found the simulator history of the pilot the most relevant. It doesn't require this long chain of relationships like some other things. It is the most impactful person recently doing something related to the event that is extremely weird behavior.

I get what you’re saying but I doubt it because air travel now is extraordinarily systematic and well understood.

Reply on news.ycombinator.com