9-11: Chaos Theory; Cockpit notification; a comparison of linear and non-linear responses


We have discussed the events of 9-11 in multiple articles using the lens of Chaos Theory and have established that events were chaotic.  Even though a chaotic situation is non-linear, the government’s habitual response on 9-11 was linear; one specific exception, Boston Center (ZBW).

It is generally understood that ZBW preempted an unworkable hijack notification protocol.  What is not well known is that ZBW also preempted the existing policy concerning cockpit notifications.  In the division of work between FAA and the carriers, cockpit notification was a carrier prerogative, a linear process.

In a recent article we established that the Aircraft Communications Addressing and Reporting System (ACARS) notification to UA 93, a linear process and a carrier prerogative, was just minutes late.

Using the lens of Chaos Theory we have established that linear approaches in a chaotic situation work poorly, don’t work at all, or are counterproductive.  Cockpit notification, by the book, did not work.  But ZBW did not follow the book, for the second time that day.

The UA 93 Notification

As I wrote in the earlier article, “United 93 received such a message at 9:23, according to the dispatcher.  Herndon Center considered such contact to be an airline prerogative and deferred to them.  United Airlines dispatchers had begun contacting pilots as early as 9:03, but not initially with specific warnings.  The first contact was to inform the pilots that aircraft had crashed into the World Trade Center.”

ZBW Notification to Cockpits

Shortly before 9:10 at least one ZBW controller took positive action.  The controller (Hampton Sector 31R) handling the Otis scramble informed a plane under his control that “we have a message for you to heighten your cockpit security due to some activity this morning.”

That transmission caught the pilot by surprise; he knew the policy.  He responded, “OK, and that is from, uh, company?”  To which the controller responded, “negative it is a general advisory to all aircraft, this morning an aircraft hit the world trade center, had been a hijack, and we are receiving reports there may be a second one.”

Listen to the  ZBW warning.

That was not the only plane so notified by the controller.  Even though his assumption was that all pilots on his frequency, to include Panta 45, the lead Otis pilot, had heard his notification he continued to check to make sure.

For example, at 9:12, he advised an aircraft that had not copied his warning that, “all FAA facilities are advising all air carriers to heighten their cockpit awareness.  There has been at least one hijack this morning and possibly two.”

Listen to the ZBW continued warning.

Note the bolded text in the preceding paragraph.  That was the policy and it is what all other air traffic control centers likely did, advise the carriers.  We know from the UA 93 case that Cleveland Center left the notification to the carriers.

I have listened multiple times to all the air traffic control tapes provided to the Commission and I don’t recall another example of what ZBW did.  Our universe of tapes included audio files responsive to our investigation from five centers, Boston, New York, Indianapolis, Cleveland and Washington; and tapes from concerned towers and TRACONS.

Management Lesson Learned

This is a personal comment.  My advice to senior government officials faced with a chaotic situation is to do two things.  First, immediately resist the temptation to follow the book; look at your non-linear options.

Second, determine as quickly as you can who is actually  fighting the battle.  If the answer is that there is more than one battle commander then make sure they are talking to each other and that information is flowing freely to and among  them.  If the answer is there is no one then immediately appoint an incident commander and demand that responsive information flow to him or her without filtering.

One final word of caution.  If anyone raises the shiboleth of “need to know,” escort them off the premises.

Addendum, February 26, 2010

The idea to make cockpit notifications may have originated with the FAA’s New England Region (ANE).  According to a transcript of the NTMO E (National Traffic Management Officer, East) position at the Herndon Center, soon after 9:09, John Bargainier from ANE attempted to pass along the idea.  He got only so far in his suggestion: “OK, the five hundered here is requesting,” before being cut off.  He was told, “everyone kind of busy…call back…in about ten minutes.”

Bargainier persisted and called the NOM line and spoke with Ellen King, who was monitoring the line for Ben Sliney.  According to the transcript of that line, which is not time marked, Bargainier said, “Bill Ellis, the five hundred has suggested, requested that ah you get an advisory out to all dispatchers for any international traffic coming in to new york (sic) to increase cockpit security.”

Bargainier repeated the request at Ellen King’s request.  “that you put a [sic] advisory out to all dispatches for a to advise that any traffic coming in to new york [sic] internationals especially to z b w [Boston Center] to ah increase their cockpit security maybe a no brainer and you’ve already done it but we just want to make sure that you know we try to do something”  King responded, “ok, I have your request.”

Note two things.  First, the mindset is to work through the carriers, their dispatchers.  As I recall, Herndon Center personnel confirmed that to us during out visits, that was the policy.  Second, note the ANE emphasis on internationals and ZBW.  ANE has just one Center under its jurisdiction, ZBW.

The ANE plea “to do something” and the existing policy, a linear process, did not translate into a timely notification to UA 93.

Addendum, March 2, 2010

Cleveland Center (ZOB) Lorain Sector acknowledged UA 93 checking in shortly after 9:25.  That exchange and immediately following controller communications provide  insight as to the situational awareness of the ZOB controller.

The controller exchanges between 9:25 and 9:26 can be heard here. 092510 UA 93 check in and controller awareness

Contrast the ZOB controller’s situational awareness with that of the ZBW controller we heard earlier.  The ZBW controller knew the specifics of the New York situation.  The ZBW controller’s awareness was that there had been an accident.  Readers are cautioned that this is just one segment of the work of a very busy controller that morning.  Interested readers who want to hear the entire tape, 2 ZOB 14 LOR-R 1319-1333 UTC AP.mp3, can access it at this link. Click on FAA and then ZOB.

Shortly after 9:30 and while UA 93 was being hijacked the Lorain controller had this conversation, an example of how busy he was.  Lorain pace of work


We wish it could have been otherwise, but Cleveland (and Indianapolis and Washington) Centers did not have the situational awareness that Boston and New York Centers had.  In part, I attribute this to the FAA’s administrative Regions, New England, Eastern, and, to a lesser degree, Great Lakes, attempting to play an operational role.  The Regions were attuned to using linear processes, the deliberate steps of accident and incident investigation.  Only Herndon Center, an operational entity buried deep in the FAA’s organization chart of the day, habitually managed chaos.  Few in FAA or anywhere else recognized that capability that day.

At one point during an important FAA HQ conference call, Herndon Center had heard enough and interrupted a situation update to tell HQ they had it wrong.  I will eventually tell that story in a separate article, it was an illuminating moment.