Hacker News new | past | comments | ask | show | jobs | submit login
Delta Goes Big, Then Goes Home (flightradar24.com)
1107 points by devy on Sept 7, 2017 | hide | past | web | favorite | 188 comments

Commercial pilot here, currently flying transatlantic routes. I avoided Irma last week while flying to Buenos Aires (it was not this big yet). I have partners in San Juan at this moment, bunkered in the hotel. We see this as a crazy move, very well executed, but too risky for the plane itself (not so much for the crew).

Let me explain, the metar and tafors (the airport meteorological reports we use) during the flight in San Juan show that the flight was doable. No crazy winds during landing, and worsening during take of but well within the limits you may find in a local thunderstorm. The predictions allowed them to see the window to fly in and out. If arriving the weather was too bad or you see the hurricane getting too close to the airport,they could have got away without landing. For take off either you have the right conditions (under the legal limits) or you don't.

What they risked IMHO is having some kind of technical problem or delay once in San Juan and having to stay there to weather the hurricane. That could mean heavy damage to the airplane while on the airport, even a total loss.

That said for me the most amazing of it all, is not the short stay at the ground. While very fast is not uncommon to see them with an experieced team, the slower part is the disembarking and embarking of the passengers, and they surely were all motivated to hurry.

The most impressive is how they followed the quite thin alley between the hurricane arms till they were clear of the bad weather. In a situation like that you only see red and magenta in your radar and a narrow black or green zone where you can fly safely. That's surely was the most scary part for them beyond the adrenalin of a fast and potentially dangerous operation.

Looks like the typical movie scene of a spaceship passing between closing doors!! Certainly brass balls.

Edit: some typos and a "don't" that I left behind and was changing the meaning of a sentence. Sorry written from the cellphone.

It's interesting that a commercial pilot would buy in to the "risk" bent of this article. Are you an inexperienced/new pilot? I flew into and out of the outer bands of Harvey last month. That portion of the system is no different than any storm. Nor is flying between the bands. That gap is huge compared to a little tiny airplane. No "adrenaline" or "balls" were involved here. Just competence. I'm sure the pilots have had far more harrowing experiences flying back and forth across the US in the summer and the departing passengers were in no danger.

Now i wonder who in Delta management were willing to put the plane, never mind the crew, on the line for this flight.

It if was totaled by the hurricane it would not have looked good on the company books.

Very good question. Whether Im stockholder or shareholder in this position I want to know why and why is risking hundreds of millions of dollars in my rolling stock not to mention lost lives and PR disaster.

Honestly there is NO reasonable explanation. But wait few days to find out there was some high ranking company member or CEO/C level exec or his/her family stuck over there and thats why they send a plane.

And no - if it wouldnt take off that minute, you are talking about wreck taken off the books.

The optimist in me would like to believe that Delta took this risk, because they considered their humanitarian mission more important that the mechanical integrity of one plane. I would encourage each and every company I own stocks in, to prioritize their goals similarly.

Every situation is going to look different. But if I found out a company I was invested in had an opportunity to save lives at a reasonable risk[1] and chose not to due to concerns about whiny shareholders and PR, I'd dump them and write comments about their shitty morals on message boards.

[1] Whatever that means. And I have no personal opinion about the risk involved in what Delta did, because I'm not qualified to form one.

Edited to add: it is nice to be able to applaud an airline for a change. Yay, Delta.

Given that lives were not in fact at significant risk, it's conceivable to me that they considered the amount of free publicity this would likely get. Given that, and the relatively small chance of serious damage or loss of the plane, and it could be a decision with positive financial expected value.

Plus, of course, there is the humanitarian benefit.

As explained by omegant in the post above, lives were not at risk here.

The risk was to the aircraft: if it had been delayed on the ground at SJU for any reason, it might not have be able to depart before the storm hit and would likely have suffered severe damage.

Normally I would think that

> hundreds of millions of dollars in my rolling stock

would refer to the aircraft ... except that's a nearly 1-order-of-magnitude overestimate of it's value (a single $100M, not multiple).

Not sure if there were really any passengers disembarking in San Juan. I would guess that it was pretty much an empty flight flying in.

What a heroic effort! Salute to the entire team!

There were 30 people on the flight to San Juan.

Each one of them must be quite a story. Being in San Juan despite the danger must have been very important for most, and they almost were unable to make it.

Airfare was cheap to make it an attractive mileage run.

But MQDs.

Get that waived with the spend on the credit card.

Perhaps disaster response folks pre-positioning. Red Cross, etc.

Or people rejoining elderly relatives or other dependents...

Honestly wondered if Jim Cantore was on that plane.

Heh, great comment. I wrote The Weather Channel a letter (yes, on paper, sent by mail) years ago specifically in appreciation of Jim's demonstrable love of severe weather. He is a guy who clearly loves his work.

This comment is a great example of why I read the comments on this site. Thanks for the answer. It provided some context I'd been wondering about.

One often neglected risk is the risk to the ramp crew from lightning. For a quick turn around in the middle of a storm like this, someone needs to neglect their own personal safety. Many times lighting is less risky to the plane but even before the storms roll in the lighting picks up.

Hurricanes tend to have far less lightning than everyday storm fronts.

Thank you for the details from a Pilot's point of view. I was watching this flight play out live and it was certainly a nail biter. I couldn't believe they were going to land let alone take off.

Amazing. Anyone have a video of the flight path with radar?

Is it possible to show the playback of the flight with weather?

seems various weather layers are included in the higher tiers of payment packages (the business package in particular).

Not sure if they are also part of recorded flights though.

Pictures of the departure with weather radar.


Thank you for the insightful explanation!

Stupid question: what about flying above the hurricane on the way back ? Possible ?

This is exactly what you'd do in a Concorde, U-2, SR-71, F-15, or F-22. All of these can get to the required FL600, at 60000 feet.

The F-15 and F-22 have an extra bonus. They can climb straight up through the eye. The U-2 could probably manage a spiral climb within the eye.

Depends how tall the storm is. Planes engines breathe air, and there is less air the higher you go, so there's a limit to how high you can go; roughly 10km for commerical jets (actually it's a factor of altitude and speed; at high altitude, you need to go at least a certain speed to get enough air into the engines, but if you go too fast the plane breaks apart). I think big storms are about 10km tall.

It's actually less about the engines and more about the wings. As altitude increases, the air becomes thinner. That means that wings stall speed increases. At the same time, as the air becomes thinner, the speed of sound decreases - and breaking the sound barrier will break your plane. As you go higher, the stall speed and the speed of sound converge.

The point where they meet is called "coffin corner". It is really not a nice place to be: If you are in coffin corner, if you try to descend, you risk speeding up and breaking the plane. If you try to slow down, you risk stalling - although it is possible to recover from a stall, you cannot recover without speeding up.

It is generally advisable stop climbing before you box yourself in like that.

What would a pilot do in a situation like that?

Descend very gently.

Did you fly Atlanta - Buenos Aires? I flew on Wednesday 6th. Were you my pilot? :)

During the cold war my grandfather flew bombers into the eyes of hurricanes to drop sonar beacons. I'm sure he'd get a kick out of this.

The older bombers they flew were all manual controls (as opposed to fly by wire), which meant physical exhaustion from wrestling the plane into the storm was a major factor - so they'd bring extra fuel to be able to circle the eye of the storm for some amount of time to build up the strength to fly back out.

This is the second such instance in the past month. - For a longer discussion on the topic see this thread on airliners.net http://www.airliners.net/forum/viewtopic.php?f=3&t=1373093

- On August 24 Typhoon Hato hit HongKong hard. All airlines canceled their service in and out of HKIA, except a lone KLM flight from Amsterdam. Similar to the delta flight into San Juan, this flight created a large media storm into the risk involved. http://liveandletsfly.boardingarea.com/2017/08/24/klm-747-ty...

While I am not a pilot, I know that such flights involve sophisticated planning (route, fuel, approach), weather prediction as well as diversionary plans if things do not work out. No airline would undertake such a high-risk move without the confidence that they are not putting lives at risk.

Consider that they were carrying 170 people away from a hurricane that killed, among the millions of people that were on its path, a total (until now) of 11. If that single plane crashed, it would have multiplied by 15 the total death toll of the storm. I'm sure that the risk must have been quite low for the operation to make any sense.

Generally, after a disaster, the death toll is artificially limited to bodies reporters can actually count, before rocketing up as infrastructure comes back online

That's not what happened during the 9/11 attacks. The death toll was reported as more than 6000. Weeks later it was suddenly adjusted downward to 3000ish and finally settled at 2997.

The extreme overestimate was ostensibly due to "double counting", which I find to be a wholly unsatisfactory explanation. With all the info that they collected about each person who was missing -- name, DOB, photos, ID numbers, where they worked in the building or who they were visiting -- how can they possibly double count 3000 people?

And you're saying that generally after a disaster it works exactly the opposite way: a low number initially which gets adjusted upwards. Interesting. Almost as if there was a desire to make 9/11 look even worse (as horrible as it was).

Wired's write-up from yesterday has a few more details including some quotes from Delta personnel: https://www.wired.com/story/delta-plane-hurricane-irma

How could neither article mention the name of the pilot? Just the company "Delta".

Because this is marketing.

Probably because the pilot didn't choose to make the flight, the pilot was just doing their job, and their name is irrelevant.

Edit: What I meant by "didn't choose to make the flight" was that the pilot didn't wake up and decide he was going to San Juan that day. He was assigned the flight, and presumably did have the authority to say it would be too dangerous. Maybe not though, I'm not an ATP.

As easy of a target as airlines are for the usual criticism, this is an heroic feat. I applaud the efforts of everyone involved.

I can't image what it must be like to fly into a storm system like that. Nerves of steel.

Hurricanes aren't really difficult to fly through since the air moves horizontally, and is mostly predictable, it doesn't really affect the flight much. Takeoff and landing are the major problem areas.

Thunderstorms, however, are much more difficult to fly through, because the air is moving vertically in an unpredictable fashion.

Yeah but what if they couldn't take off in time? Those pilots/crew would suddenly be stuck on an island weathering one of the most powerful hurricanes in history.

Recorded history.

History already means recorded. Anything before records exist is called prehistoric.

The problem with that definition is that "prehistoric" is a lot more recent than people realize in the Americas.

Interestingly the Klamath people have an oral history of the eruption of Mount Mazama and formation of Crater Lake over 7000 years ago.

I wouldn't trust that one bit. Have you ever played telephone? 7000 years is a lot of generations, it's likely completely inaccurate.

That's incredibly interesting. Does it line up with the geographical record?

Just a few days ago there was a village in Canada discovered thanks to oral tradition. It has been dated thousands of years older than the pyramids

No, "prehistoric" is the time period before written records. If we only started paying attention to hurricanes in the gulf 200 years ago, that doesn't mean 1805 was prehistoric.

S/he didn't make this up. History is the study of the past trough written records.

The word has been since overloaded to be the study of the past. Which is fine, except that leaves no reason for the word pre-history

You're partly right. However, it's very common, even when people want to draw exactly the distinction you're talking about, that they'll just use "history" to refer to a period for which there are any records. That's a pretty different concept from the period for which there are records of the strength of hurricanes.

This raises an interesting point. Do you know how long we've been accurately recording cyclones?

The National Hurricane Center has some records [1] going back to 1851 [2]. I'm sure those are based more off eye witness accounts.

[1] http://www.nhc.noaa.gov/data/#tracks_all

[2] http://www.nhc.noaa.gov/data/tracks/tracks-at-1851.png

Depends on what you're trying to measure! Arguably we still can't measure many things accurately.

I wonder if the pilots volunteer for the run? It'd be interesting to know the internal culture with regards to it.

The captain makes the same call on every flight. The plane and passangers are her responsibility regardless of the situation.

Edit - also the call is made continuously. They can back out any time.

Why do you put "her responsibility", instead of "his responsibility", as captain's grammatical gender is masculine?

English doesn't generally have gendered nouns, and the few it has are generally going out of use. I'm not aware of captain as being gendered in English.

I grew up speaking english, including reading a few older books, and I'm not aware of any gendered nouns at all. Which ones are you thinking of?

I was thinking of pairs such as steward/stewardess, actor/actress, and fiancé/fiancée. They're not gendered in the sense of nouns in French or German, but are gender specific.

How do you choose between his or her in this case?

This is more involved than I care to get into here. Here's a Wikipedia article which discusses it:


Why create an account to make your initial comment? It doesn't seem very substantial or on-topic. The two reasons I can think of aren't very charitable, which is why I'm asking, assuming that it's neither of those.

My charitable explanation is that it's a non native speaker of English genuinely asking the question to learn.

You don't, unless it's to create off topic arguments/trolling.

To answer your question if you know the sex, you say the sex.

If not you use sexless pronouns like "their" and "they" even if they are not 'proper' by old strict standards, basically they are, because language evolves.

They did, this was a volunteer crew. Most of the passengers on the return flight weren't passengers in the normal sense of the word, but non-reqs, delta and other airline employees that were working in that airport and wanted to get out of dodge.

Who was on the arrival flight?

Mostly officials and emergency personal, according to other reports.


It'd be really nice if we could leave the reddit vernacular over there.

>> I wonder if the pilots volunteer for the run?

They may or may not have volunteered for the flight, but they do get the final decision to go in or turn back once the airline OKs it.

That is why they were able to charge more than 3k per ticket to fly out of troubled locations on Irma's path.

Two points:

1) Hurricanes are actually harmless to planes. The only part of the flight that includes any danger is landing, due to wind and low visibility. But flying at cruising altitude or even through the storm is fairly harmless, as most winds are horizontal and nothing out of the ordinary for a plane. Keep in mind that the NOAA regularly flies commercial jets (such as a Gulfstream V) into hurricanes. Also keep in mind that never in the history of aviation has a plane been brought down by a cyclone. Thunderstorms, on the other hand, are much more dangerous. The main difference is that in a cyclone, air moves horizontally, while in a thunderstorm it moves vertically in bursts, which is much more difficult to navigate.

2) From what I've heard, this was not a commercial flight but a flight mainly aimed at evacuating non-rev airline employees. The main issue with doing this flight for regular passengers would be fear and discomfort due to turbulence. Airline employees on the other hand are used to flying, know that turbulence is harmless, and would rather just get home.

As someone who has lived through some of the strongest hurricanes in recorded history you are wrong on multiple counts. Flying through the storm and the wall eye is dangerous and NOAA has lost planes to it. One of their hardened P3 had equipment and life raft ripped from their anchors [0]. There is definitely lightning and waterspouts during a hurricane. That is not reserved for thunderstorms.

This was a flight to evacuate crew who are stationed in PR. Non-rev refers to crew who are commuting or on holiday. On a typical flight a plane may turn over in a few hours but the crew may remain for 24 hours and leave the following flight. This crew was down route and needed to head back to their operating base.

[0] https://www.wunderground.com/resources/education/hugo8.asp

Non-rev refers to any non-revenue passengers. You're conflating NRSA (non revenue space available) with non-rev generally. There's also NRPS (non revenue positive space) and "must ride" flyers in that category. This flight was picking up airline personnel, not revenue passengers.

Your having lived through hurricanes on the ground is entirely irrelevant to a discussion about the safety of overflying the hurricane.

At the flight altitudes of a commercial jet none of the effects you mention happen. You need to plan your path carefully for the descent, landing, and takeoff. But at cruising altitude, hurricane winds are not a major safety concern.

> flying at cruising altitude or even through the storm is fairly harmless,

Yes, so harmless that no one does it. NOAA has to use an extremely reinforced, highly maintained fleet of P3 Orions. You've mentioned repeatedly in your parent posts that hurricanes are harmless to planes when commercial pilots in this forum say they wouldn't do it without excellent weather planning.

> You need to plan your path carefully for the descent, landing, and takeoff. But at cruising altitude, hurricane winds are not a major safety concern.

Planes have all sorts of cruising envelopes. Gulfstreams are in the 40k feet. Commercial jets cruise at 25-35K foot band. And commercial turboprops (commuter shuttles) are in the 10-20k foot band. While the first 2 won't feel the gale force winds they can experience the updrafts as turbulence. Commercial jets fly in the lower stratum of the stratosphere. There is still weather phenomena at the height, just much less than the troposphere.

Finally, the captain will avoid a hurricane even if it doesn't really affect them. They may need to decend to a lower attitude for an engine flameout, for example. And why deal with the added complication of emergency landing during a hurricane.

There are plenty of ancillary reasons to avoid overflying hurricanes when possible—such as passenger comfort, to avoid possible turbulence; or if it disables nearby diversion points in case of an issue inflight.

But people here are making wild claims that overflying hurricanes is inherently dangerous and puts lives at risk. That's factually incorrect and is what I came here to correct. You (and others) are making examples about the NOAA's Orions (which are reinforced to fly through, not over hurricanes, flying into the wildest parts on purpose), and or other turboprops which fly 10-20k ft. That's not what we're talking about here… we're talking about commercial jet aircraft which fly at 30-40k ft. Business jets fly even higher, sometimes up to 52,000ft. Just this week, I flew private in a Citation X at 49,000ft.

Hurricanes reach to 60,000 feet.

It appears from this Quora question[0] that it it is, in fact, not dangerous to fly OVER a hurricane. The keyword is over.

[0] https://www.quora.com/Is-it-dangerous-to-fly-over-a-hurrican...

NOAA does not fly their Gulfstream into hurricanes, it flies over them. They use WP-3s (strengthened P-3 Orions) to fly through the hurricane. The wind shear and turbulence can put quite a lot of stress on the airframe, it's nowhere near harmless. Crosswinds also make flying difficult, not to mention the vibration. A non-reinforced plane with a good pilot and 3-point seat belts could probably make it through, but the extra maintenance needed after would be enormous.

1) No they're not harmless, updrafts and downdrafts exist


Ok, they're not entirely harmless (e.g. if you fly into an hailing embedded thunderstorm at low altitude). But overflying one at cruising altitude is 100% safe.

E.g. the plane in this story flew through the eyewall of a hurricane at 1,500 feet. A commercial jet flies at a cruising altitude of up to 45,000 feet.

Do you have any source for your claims?

The planes regularly face 300 MPH+ winds in normal flight because of their speed.

That's not evidence that an additional 180MPH would be harmless. For one thing, those 300MPH winds are normally only in one direction.

The plane moves relative to the wind it is in. A 180mph cross wind, head wind, or tail wind only affects its ground speed.

I've been on flights from Asia to the US with a 150+ mph tailwind in the jet stream making the ground speed 700mph+. However, from the perspective of the air frame, it's still just traveling at a normal speed relative to the air it's in.

So I'll take that as a "no" that you don't have any citations or professional/academic experience to talk about airframe stress dynamics in a hurricane.

What citations would you like beyond basic reasoning skills? Planes don't fly relative to the ground, they fly relative to the air. Unless the wind is gusting from 0 to 160 (which is not how hurricanes work), there is nothing to be discussed here.

The earth rotates at 1000 MPH and you don't see people wondering if cars can handle that.


Because air moves in more than one direction. Wind shear is a serious threat to passenger aircraft.

Thanks for the context. The other comments make it feel like this was extremely risky. Like "moving spacecrafts through closing doors".

You give it more context

But they give ZERO source for their claims.

It seems like you're looking for a particular point of view here.

Yeah you might be right, I just took it at face value. Also, I'm pretty unaware of the specific realities in the interaction between planes and the weather.

Thanks for pointing this out though

Does anybody who works in the industry have any idea of what the risk management is like for these types of events?

What are the odds something could go wrong compared to the relative good of squeezing in one last flight's revenue for a while? Is flying in a hurricane actually not that bad and the airport turnaround/takeoff the most dangerous part?

I'm usually cynical in these situations, but I don't think it was about revenue. Quick decision making by even the most incompetent executive would have prompted the following response:

"Let's see... 50% chance of incremental $200*150=$30K (idk the price nor the capacity, so just entertain me) in revenue. Let's bump that with the expected loss/lawsuits if this fails = ... Who cares? It's in the millions. Just don't do it."

I think this is one of those cases where the airline took it upon themselves to really try to help as many people get out as they could.

It's also my understanding that the passengers they flew in were FEMA workers and 100 or so of the people they flew out were "nonrevs" with seats still open on the plane after that i.e. Most of the people who they flew out worked for airlines and no one was kept off the plane due to space

For reference, the cost of a 737-900ER is about $100M. The value of the PR that comes from a heroic evacuation is hard to measure.

How about the cost of PR if the plane crashes and your airline looks insanely reckless and guilty of manslaughter?

Even if the plane crashes, a properly planned flight shouldn't look reckless - knowing the risks, but deciding them manageable and the best option for everyone involved doesn't seem too far removed from a normal flight.

It's also not just an airline decision - you've got a pilot in control who can make the no-go call at any point, and their training should prevent them from undertaking unnecessary risks.

Your analysis is rational but the discussion is about public relations outcomes, not decision making. The public response to issues like this one is almost always not rational and does not rely on risk management.

For trying to evacuate people in front of a life-threatening storm? That's not reckless. (Granted, the media might paint it that way...)

> I don't think it was about revenue

Just look at all this free PR.

But I choose to believe it was about giving people one last chance to get out.

PR benefits seems like another plausible explanation.

But the way you imagine it makes it sound like a complete coincidence that they happened to time it so perfectly well so as to use the gap of calm air effectively... whereas that surely seems pretty well-calculated.

I don't quite follow what you're trying to say. My comment makes no suggestion of any coincidences, nor mention any timing to use calm air effectively.

Yes, timing calm air effectively requires calculation. I'm at a loss when deciphering how that relates to what I'm saying.

What I'm saying is, you're saying this was a case of the airline trying to get out as many people as they could, without much regard for the risk and costs. But from what I see, they seemed to have optimized for the risk and the cost extremely well, to the point where it made perfect sense to do what they did, both business wise and humanitarian. Meaning those seem contradictory.

(In no way am I playing down what they did though. I'm only commenting on your analysis.)

There is such a thing as being heroic & not stupid.

I didn't say otherwise? Are you responding to my actual comment, or something else?

I'm sure they did their best to calculate it as precise as they can. Other airlines tried and failed. They had to turn back or were stuck on the island.

So the realized risk is fuel and cost of a grounded plane.

They could have lost the plane according to the Wired article:

> Once the plane landed in San Juan, the alternative—leaving the plane on the tarmac or in a hangar—would probably have destroyed the airplane ... If the plane had gotten stuck, Delta probably could have gotten the humans out safely. But it almost certainly would have lost a plane.

Hurricanes are relatively slow-moving, relatively predictable, and have relatively consistent wind speeds/directions.

Obviously it would be quite bad to fly into a hurricane's rain bands (the weather outside of the "corridor" that the DL431 flight moved through).

The real feat here is the real-time and reliable coordination between all interested parties -- as mentioned in the article, the meteorologists, dispatchers, pilots, cabin crew, and ground crew. Everybody involved had a distinct sense that the window of opportunity was narrow and it should be quite clear that they acted with a high degree of professionalism and skill.

As for the life-threatening danger: it's risky, that's for sure, and I would by no means calls this a small feat, but there is no intense wind shear like those found in tornadoes, thunder storms, microbursts, or other such things.

So you are saying thunderstorms contain more intense wind shear than a hurricane...I don't think I believe that.

No, he's saying that the wind shear they had to contend with in the calmer corridor between rain bands was less intense than that found in a thunderstorm.

I am reminded of this story I bumped into on Reddit some time ago. Granted it's a small plane but still:


The blog indicated that it was an evacuation flight. It may have been non-revenue.

Flying through a hurricane sounds dangerous, but it's actually not much different than flying normally and is perfectly safe. Planes are very resistant to horizontal wind, which is all a hurricane is.

You might be overselling your point with "perfectly safe". Extreme vertical winds are not uncommon in hurricanes. See

With some quick Googling, I get the sense that while people cite hurricane hunters, they gloss over important bits,

  But what does get interesting is flying through the 
  hurricane's rainbands and the eyewall.

  -- http://www.aoml.noaa.gov/hrd/tcfaq/H3.html
In particular, rainbands occur all throughout a hurricane. You can't really fly through a hurricane and not fly through these. I'm not an aviator but I grew up in Florida and have experienced several hurricanes. I get your point about hurricanes being mostly horizontal winds and why thunderstorms are much more dangerous. (I appreciate your post because it motivated me to learn more :)

But hurricanes are immense, and immensely powerful. The size and energy drive all manner of dangerous phenomena. For example, growing up in Florida we were worried less about the hurricane, per se, than the many tornados they often spawned.

Flying into a hurricane involves a lot more than just horizontal wind. There's convection, turbulence, wind shear and heavy precipitation. It is not uncommon for hurricanes to spawn tornadoes.

NOAA has been flying aircraft into hurricanes since the 1950s. It is not inordinately dangerous, but it is not just another flight, either.


Take off and landing are especially dangerous in a hurricane; where wind gusts can exceed a difference of 50 knots.

Does ethics play a decision here too? As in "let's try our best to get these people out of here". Or is the risk actually the opposite, and it would be safer for them to stay in Puerto Rico?

Wow, this is really incredible. The flight path out of Puerto Rico is nearly perfectly optimal to avoid the weather.

I wonder, is this a good (and incredibly positive) example of what quality real-time/near-real-time data can do for decision-making?


I wouldn't glorify the analytics too much in this case. I'm confident that the management and the pilots made decisions that far exceeded decisions that could be made with data approximations.

Can't make decisions without data. Sure, management and flight crew deserve the lion's share, but that real-time data is probably critical for those decisions.

You know those planes have their own weather radar right?

Based on what?

Made decisions based on what? Astrological charts?

You need meaningful data to make meaningful decisions. Otherwise you are mistaking luck for expertise.

> I wonder, is this a good (and incredibly positive) example of what quality real-time/near-real-time data can do for decision-making?

Yes. Obviously.

I guess what I was trying to say was, what planning was done beyond the heroic measures and decisions made by the pilots and supporting infrastructure people? I was curious as to whether this was pure bravery with an informed decision support based on data, or simply pure bravery with a little bit of luck. I am always trying to find positive examples of how data is used for good, to save lives, etc. The more we know about how we can use data for good, the more we can try to do good with data...

It's hard to say without knowing more details, like exactly what kind of avionics the plane had. Ironically, the avionics in many commercial aircraft are quite dated. The plane I fly, a Cirrus SR22, has more advanced avionics than many commercial jets. But my guess would be that that the plane was never in any real danger, and the main risk was that it could take longer than planned to turn the plane around on the ground. Then they could be stuck because the storm had moved in and they couldn't take off any more. For example, if they had blown a tire on landing, or developed any kind of engine trouble, they probably would have been screwed (where "screwed" means stuck on the ground in San Juan, not crashed). That's the sort of thing I would have been sweating about if I had been making that flight. Otherwise it was really business as usual -- the extreme end of BAU to be sure, but still BAU. Pilots have to fly around heavy weather all the time.

But yes, this flight would have been absolutely impossible to conduct safely without satellite and weather radar data, as well as good forecast models of the storm's movement.

I hate the title because it sounds like a failure when in fact it is an excellent thing Delta did.

I agree, the pun has a negative connotation whereas the article is about a feat.

I'm wondering about the ground crew willing to stay in the airport to help this flight get going - and then having to get to a shelter in the middle of a hurricane.

While I'm not familiar with SJU, my guess is an airport terminal is likely one of the more sturdy buildings around, and is likely to have some pretty robust structures you can take shelter in. Most commercial buildings are sturdier than the average house, and I'd rather be at an airport than at home in a storm that levels buildings.

I used to live about 20 minutes drive away from SJU (my parents still do) and travelled through it frequently for college breaks. SJU is much like any other largish international airport in the U.S.

Many of the houses and buildings there are sturdier than you'd think. My parents house is at the end of a row of six or so that share walls. So they buffer each other's flanks during a hurricane. Their windows don't have glass. Instead it's screens covered by storm shutters with iron bars over those. Also, the houses are constructed of thick reinforced concrete rather than the picturesque wood, siding, and shingles that you might be thinking of. It's actually rather bunker like. Hurricanes can hit several times per year and those houses are built to take the punishment.

(Got word from my parents last night that the power was out and their street is filled with windblown debris that they and their neighbors will be cleaning up today, but otherwise they came through fine.)

I dont think the fear is sheltering... Most of those islands have shelter, its not like hurricanes are a rare event for them.

However, its the destruction and risk that it could be WEEKS before help comes if infrastructure is completely destroyed... and an island with limited resources means that the people still on the island will (compared to mainland) very quickly start to realize that they need to eat and acquire things to survive... and survival mode is the worst part of any natural disaster. The damage done by Katrina to New Orleans was horrible, but honestly, the damage done afterwards by the people that lived there (that was done to survive, im not criticizing, morality changes real quick in survival mode) wsa pretty horrible, too. Imagine a small-ish island with not a lot of natural resources and no quick way to get truckloads of supplies en mass...

Not saying thats what's going to happen - but thats the kind of outcome people forget when they volunteer to ride out a natural disaster. The actual event is just the prelude to a pretty bad period of time for the area it hit... and the pain you will feel after the hurricane is likely to be much higher than the hurricane itself.

And this is not just any old commercial building, but is also probably considered vital infrastructure for emergency planning, and so extra-reinforced.

I'm glad someone bought this up, it's the first thing that came to mind for me. Even if they are safe in the airport, I'm sure there were more important things they could be doing like securing their home or buying food and water.

The downside of the internet... suddenly everyone pretends to be an aviation expert.

And no IANAP disclaimers

I am surprised this flight was approved and permitted by airline, government, and insurer procedures... the risk/reward looks exceptionally poor, as the storm could amplify a normally-moderate, manageable mechanical failure into a situation that greatly overwhelms the flight crew (and there is definitely precedent for that in the accident record).

There are probably at least some NTSB human factors specialists cringing at this.

Other than a very bumpy ride and lots of nerves, I would think a 737 (and many other commercial jets) would have no problem flying through a hurricane. Landing on the other hand...

Anyone with knowledge of this?

Private pilot with 20 yrs experience here...

To say that a 737 would have "no problem" would be overly optimistic. A 737, like all commercial airliners, is pretty tough, but there are plenty of things in a hurricane that could bring one down. There's hail, up-and-downdrafts that far exceed the capability of the aircraft to overcome them, and turbulence severe enough to flip the plane inverted. The least likely event is turbulence severe enough to structurally damage the plane, though even that is possible, and hail could definitely damage the plane, especially the engines, to the point where it would no longer be airworthy. You might make it, but you'd be rolling some pretty serious dice.

Landing in a hurricane would be impossible. You can handle being kicked around in the air because there's nothing to hit. But near the ground you need reasonably stable air in order to make a controlled landing. Anything more than around 50 knots is pretty much a show-stopper for any civilian aircraft, and even much slower winds than that can present significant challenges if they're gusty or not aligned with the runway.

A hurricane isn't a thunderstorm, and has very limited vertical movement in its airflows. The hurricane hunter P3s do it.

> A hurricane isn't a thunderstorm,

That's true, but they can (and usually do) contain embedded thunderstorms.

> and has very limited vertical movement in its airflows.

Except for the embedded thunderstorms.

> The hurricane hunter P3s do it.

That's true, but they can do it because their destination is the eye of the storm. That allows you to do all your navigation relative to the storm.

Commercial airliners are trying to reach a destination on the ground, i.e. a target that is moving relative to the hurricane. Trying to navigate with this additional constraint is a whole different ballgame.

Also the hurricane hunter WP3s (and the USAF equivalent WC130s) all have 3-point harnesses (or better) for all seats. That's pretty important for passenger safety in turbulence. Passenger airplane lap belts leave a lot to be desired when dealing with strong turbulence.

This is what you get when an airliner goes into a big storm. http://www.telegraph.co.uk/travel/news/pilots-awarded-medal-...

I'd think that flying through one would be pretty doable. I don't think the "hurricane hunter" planes are anything special in terms of their ability to withstand turbulence. You may need to avoid the parts with the most severe convection. Airliners don't necessarily fare well in the middle of violent non-hurricane thunderstorms either.

Landing and taking off in one, however, would be really bad once the winds pick up.

It's not the same at all, but yesterday my flight had 150mph head winds (compare to Irma's 185) and you couldn't even notice. I only did because of the news coverage and wanted to compare.

Flying through a hurricane is not dangerous at all, especially for a massive plane like a 737. Airplanes are incredibly resistant to horizontal winds, which are all a hurricane is.

It's vertical winds that are dangerous to fly through, which is what thunderstorms create.

Hurricanes usually have embedded thunderstorms.

So does non-hurricane air. The hurricane by itself presents very little danger to a plane in the air.

> So does non-hurricane air

That depends on what you mean by "non-hurricane air." An embedded thunderstorm by definition has to be embedded in some other weather system. That other weather system doesn't have to be a hurricane, but so what? That doesn't change the fact that one of the many dangers of flying into a hurricane is the embedded thunderstorms it contains.


"Hurricanes and tropical storms can also produce tornadoes. These tornadoes most often occur in thunderstorms embedded in rain bands well away from the center of the hurricane; however, they can also occur near the eyewall."

> The hurricane by itself presents very little danger to a plane in the air

That's a vacuous statement. Yes, if you remove all the dangerous parts of a hurricane, whatever remains is, by definition, not dangerous. And yes, it's true that strong winds by themselves are not dangerous. Commercial airliners regularly fly through the jet stream which is comparable to a hurricane in terms of wind speed. But that doesn't change the fact that hurricanes contain embedded thunderstorms, which can be very dangerous.

The best part is that the hurricane's tailwind shaved 40-odd minutes off the itinerary.

Most of the early return happened because the ground crew managed a really quick turnaround and so the flight departed Puerto Rico 25-30 mins early.

To me the hurricane tailwind (blowing South) appeared to be against the direction of flight (heading North)... but the flight arrived early so I must not be parsing that right.

the upper level winds are actually a high pressure area and thus flow clockwise, giving it a northerly boost at that level

I'm curious about who was coming in on this flight, and why? Was it mostly people who wanted to ride out the hurricane with families? First responders?

That's pretty awesome. Does anyone know more about this type of pre-hurricane flight? Is that a common thing airlines do, to try to squeeze one last flight in? Are there rushes before other types of bad weather? And do they increase (or lower?) prices in these cases?

Other airlines tried the same thing (https://twitter.com/AirlineFlyer/status/905433106475311104) but had to turn back.

This is an impressive team effort by Delta--and a great demonstration of the teamwork required between those in the air and on the ground to pull something like this off.

why? the hurricane is predictable in its movement and moves at a snail's pace relative to the plane. The pilots I'm sure had radar data inside the cockpit

It's not rocket science.

The plane was fully turned around and refueled in 52 minutes, with a hurricane bearing down on it. Your response doesn't do a lot of justice to an impressive team effort.

the only plane at the airport right? full dedication of available personnel

You must be a professional pilot. Yes, I bet that's it. Plus, flying into and out of hurricanes isn't rocket science (and everybody knows that rocket science is really difficult and dangerous), so, easy stuff.

> You must be a professional pilot. Yes, I bet that's it.

I won't claim one way or the other, but let's presume that this person is a professional pilot. Does that not make their opinion/analysis more valuable, rather than explaining it away?

There is much to touch on from the comments so far, but I'll try my best to keep it on point. (Note that I didn't say "short" <wink>.)

Bona fides: FAA Licenced Aircraft Dispatcher; 11 years industry experience. (Left the industry in 2000) Staff title: Chief Dispatcher

Preface: I'll not attempt to address the many meteorological or airframe engineering aspects of this mission other than to note that Delta staffs its own meteorology department (or did last I was privy to their operations). The capabilities that lend to the carrier are net positive, as should be obvious.

On to some specific questions raised ...

(0) >>>[qume] The captain makes the same call on every flight. The plane and passengers are her responsibility regardless of the situation. Edit - also the call is made continuously. They can back out any time.

Response: Only half true: the pilot-in-command (PIC) along with the aircraft dispatcher share responsibility for the "initiation, operation and termination of the flight." (Yes, I think the regs use 'termination'; that always made me wince.)

So, it's a quorum of two: if either one chooses to terminate (or 'not initiate') a given flight, it cannot be operated. That doesn't mean that a dispatcher who disagrees won't spend some effort presenting evidence for his position (e.g., ten-minute phone calls), but at the end of the day of those two don't agree, the flight cannot operate.

(1) >>> [phkahler] They may or may not have volunteered for the flight, but they do get the final decision to go in or turn back once the airline OKs it.

Response: This statement begs the question: who is "the airline?" The relevant US regulations (CFR 14 Part 121) refer throughout to this entity as the 'certificate holder' -- because an air carrier holds an operator's license -- a business license, of sorts -- granted by 'the administrator,' the regulatory term for the FAA. I'd gamble that since the FAA descended from the CAA that the policy authors thought it wise to anonymise the parties wherever possible in the event of future language changes. Smart move.

So for any given flight, "the airline" would be the dispatcher, who by proxy, exercises the right of 'the certificate holder' to "operate a particular flight over a specific route under specific conditions" (going from memory, mind you).

This authorization is formally granted by way of a legal document prepared by 'the certificate holder' (read: dispatcher) known as the 'dispatch release,' which includes a minimum set of specific information (flight plan, equipment type and number, flight crew, fuel min/max/burn, alternate airports, and so forth) but typically have an abundance of supplementary information to better brief the flight crew of the expected conditions and details of alternatives that are likely to be available if the proposed plan cannot be followed.

Bonus fact: If you ever were waiting after boarding and the crew came over the PA to say they are "waiting on paperwork," most of the time it is a bag/weight count, but some times it's the dispatch release. If you know your flight is going across, or into, some crappy weather, the chances of the latter are greater than average.

(2) >>>[passivepinetree] Does anybody who works in the industry have any idea of what the risk management is like for these types of events?

Response: The airline operations hundreds of flights a day; all have some risk, and all decisions must be made in real-time. In cases of long-running events such as a hurricane, there is likely to be some general tone taken by the carrier at the highest operational levels (chief pilots, chief dispatchers, VP of Operations, etc.). For my carrier, these strategic positions would be discussed in the morning meeting, which was a recurring conference call between all those parties and department supervisors -- kind of like a stand-up, except we were all sitting around a conference table.

Aside from that, as the day wears on the specific handling of a given route, weather event, etc., is handled in real-time by the assigned dispatcher and support team (meteorologists, mechanics, etc.)

Of note: some larger carriers maintain a 'Trouble Desk' staffed by dispatchers who are assigned a lighter workload than the regular line folks. This is a great system (one that my carrier didn't have) because, let me tell you, just one fubar flight can monopolize all your capability and time for quite a while. If/when one of those flights pops off the queue it can wreck your throughput for the remainder of the shift. For my money, the trouble desk is an excellent mitigation tactic capable of keeping the workload in the dispatch office sharded appropriately.

(3) >>>[joemi] Does anyone know more about this type of pre-hurricane flight? Is that a common thing airlines do, to try to squeeze one last flight in? Are there rushes before other types of bad weather?

Response: I only had a handful of duty experiences with hurricane landings that were in our region of operation, but I do have a wealth of experience with other severe weather systems here in the US -- aka, tornado season.

There were many occasions when a strong cold front would be bearing down on cities we served, with solid lines of thunderstorms sweeping through the region. The kinds of weather that serve up severe or extreme turbulence, large hail, and tornadic activity. And there aren't any "holes" to "slip through," either.

Frequently, it would come down to trying to get one more flight in and out of a city. (The pressures of 'completion factor' at an airline are a whole discussion in itself.) In these cases, I'd be measuring the relative velocity of the line versus the distance to the airfield, estimating the time of impact, so to speak, and cross-referencing that against my computed time en route and considering the turn-time at the station, etc., etc.

Assuming you judge that it can be operated safely, and you can provide a suitable alternate plan (and "turn back to base" is certainly a common choice) you have to get on the phone (or radio) and brief the PIC on some or all of the details (all of the details are in the dispatch, but people like to hear a human voice when facing stressful situations; think 911 operators), and if they concur (or accede, in some cases) then, from (1), by the necessary joint agreement, the flight is initiated.

And as you might expect, sometimes the flight got in and out, and sometimes it diverted or came back to base. In either case, occasionally the crew might call or radio back in with reports on the conditions (or vociferous complaints about the ride quality -- hey, it happens).

So then my unqualified answer to this question from joemi is "Yes."


Finally I'd like to call out the commenters who made mention of the ground (and other station) crew and their exposure to risk in these "irregular operations"; I'd say they exhibited a measure of aplomb no less than the flight crew, and those employees are too often overlooked as essential parts of the carrier's operations; both day-to-day and in extreme cases such as a hurricane landfall.

HTH, /Acey

I'd like to think all the staff were just blasting Death Grips as they flew in.

We watched these flights as they we happening and I was on edge just watching. I can't imagine what it'd feel like having been on one.

Oh man, that is really amazing.

I guess a mix of heroes and craziness.

Congratulation to everyone involved and thankfully everything went alright.

Big balls for whole team to make this work IMO and kudos to Delta management for making it happen.

Great story.

I guess it cancels this out: https://i.redd.it/s8e8jf5gkhkz.jpg

The only disappointment for me is that they flew a relatively modest 737 down there. If you're going to make a heroic flight like this do it in an A380 or a 747 and load up as many people as you can to get them out of dodge. I guess shuffling a big plane like that at the last minute is probably not feasible since they have their own routes to service.

An evac flight could even go full on quick load. No checked bags, one carry-on max, forget assigned seats and just fill the plane from back to front as fast as possible to get out of dodge.

An A380 or 747 is significantly less nimble than a 737, and would take a lot longer to unload and load back up. The flight in question did a turnaround in 51 minutes, including taxing time, which is extremely quick.

It is, but I suspect you can significantly speed up loading if you skip all of the bullshit that makes loading take forever.

No people loaded down with multiple huge carryons, no checked luggage, no searching for your assigned seat, no zone loading bullshit, just everybody walks on the plane and sits down in the first available seats they find leaving no empty seats. Everybody lined up at the door ready to go the minute the plane is empty. You could probably have the whole plane loaded before it finished refueling.

Haha, Ryanair will beat that turnaround time in normal operations any day

The above comment is not a snide joke at Ryanair, here's an article citing their turnaround times from the early 2000s: http://news.bbc.co.uk/1/hi/business/2962698.stm

This is still the case.

Ryanair's been criticized for safety issues over their 25-30 minute turnaround times, which don't count taxiing time.

They had empty seats on the flight out. I'm pretty sure a 747-400 or A380 would have been wasted fuel and increased risk (due to size).

I have no insight into what kinds of planes are normally flown in and out of San Juan, but the plane has to be able to take off/land on the runway in question, the facility has to be able to service (maintain, fuel, provide jetways/stair cars for, etc) the plane, .. you don't (typically) just pick a random plane, fly it into a random place, and fly it out.

Why can't they use some sane colormaps instead of this garbage?

You mean the meteorological charts? It is a sane color map? What would you like to see changed or had trouble reading?

I think he is talking about the actual false colors being used. I also thought the same thing.

The picture seems to use the "Jet" Matlab colormap, which is pretty bad by many accounts:




Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact