Near midair over MSP

The National Transportation Safety Board has issued this news release on what it classifies as a “near midair” over Minneapolis St. Paul.

The National Transportation Safety Board is investigating a near midair collision between a commercial jetliner and a small cargo aircraft that came within an estimated 50 to 100 feet of colliding near the Minneapolis-St. Paul Airport

On September 16, 2010, about 6:49 a.m. CDT, US Airways flight 1848 (AWE 1848), an Airbus 320, was cleared for takeoff on runway 30R en route to Philadelphia, Pennsylvania, carrying five crewmembers and 90 passengers.

At the same time, Bemidji Aviation Services flight 46 (BMJ46), a Beech 99 cargo flight with only the pilot aboard, was cleared for takeoff on runway 30L en route to La Crosse,

Wisconsin. Weather conditions at the time were reported as a 900-foot ceiling and 10 miles visibility below the clouds.

Immediately after departure, the tower instructed the US Airways crew to turn left and head west, causing the flight to cross paths with the cargo aircraft approximately one-

half mile past the end of runway 30L. Neither pilot saw the other aircraft because they were in the clouds, although the captain of the US Airways flight reported hearing the Beech 99 pass nearby. Estimates based on recorded radar data indicate that the two aircraft had 50 to 100 feet of vertical separation as they passed each other approximately 1500 feet above the ground.

The US Airways aircraft was equipped with a Traffic Alert and Collision Avoidance System (TCAS) that issued climb instructions to the crew to avert collision. The Beech 99 was not equipped with TCAS and the pilot was unaware of the proximity of the Airbus. There were no reports of damage or injuries as a result of the incident.

NTSB and FAA investigators conducted a preliminary investigation at the Minneapolis airport traffic control tower on September 18th and 19th and are continuing to review

the circumstances of this incident.

In the past, when I’ve forwarded these reports of near mishaps, some pilots have suggested it’s much ado about nothing. This is different. Fifty-to-100 feet in the clouds? That’s a big deal.

To help you visualize these things, both planes took off on parallel runways, heading in the same direction. That happens all the time. Turning one plane into the path of another is highly unusual.

Update 11:55 a.m. – Here’s the audio of the conversations that morning. The controller ordered the left turn — to the south — for the Bemidji flight as he gave the flight permission to take off. Normally, that turn would begin when about 500 feet off the ground, probably before the end of the runway. The turn would send the plane away from the parallel runway, where the US Air jet was also taking off.

  1. Listen Featured Audio

A few minutes later, the controller asks the Bemidji flight if he’s “in the turn.” The pilot doesn’t understand the question and asks for it to be repeated. It’s not repeated (see note below). A minute or so later, the pilot asks to change frequencies to the departure controller and is granted the request. From the sound of things, that happened after the near miss. The controller asks, “why didn’t you start the turn after departure?” The pilot’s radio is nearly unintelligible.

The clearance and any conversation with the US Air plane is not on the tower tape (which I’ve edited and telescoped), raising the possibility that it was occurring on another frequency. I’ll post that frequency tape in a moment.

Update 12:17 p.m. – As suspected, there were two planes on two different frequencies here. Here’s the tape of the “departure frequency” when the US Air pilot (Known as “Cactus” because it’s an Air West flight operating under the US Air colors) reports the near miss. The controller says he thought the Bemidji flight was going to go straight.

  1. Listen Featured Audio

As with most disasters — and near disasters — this looks like the typical “chain of events,” the breaking of any one of which — repeating a question, repeating an instruction, knowing what each plan was for each airplane — would’ve prevented it.

Of course, an investigation will take place, but this one isn’t going to be hard to figure out. (Audio via

Update 3:08 p.m. I’ll be on All Things Considered tonight to talk about this and Steven John asked me what could have prevented this. I don’t have all the facts on this case so it’s informed speculation at best. But as a pilot, there are a couple of things I noticed:

On the tower tape, I did not hear either a request from the LaCrosse-bound flight to change to the departure controller’s frequency or an instruction to change to departure frequency. I don’t know if that’s even required (although I believe it is, I don’t have my rulebook at work). But the tower controller asked two minutes after the La Crosse-bound flight took off whether the pilot had made the turn? That would indicate that the controller knew the guy was still on his frequency, wouldn’t it?

The Bemidji pilot did not repeat the full clearance he got to take off, but he wasn’t required to. Some of us private pilots like to repeat the whole thing (including the instruction to turn left), so that a controller can pick up on our mistake and correct it before it’s a problem.

The departure frequency tape indicates another possible problem. This incident occurred right at that moment when a pilot makes a transition from tower to departure. In fact, as you can hear, the US Air pilot asks “what’s this guy doing off our left” before the departure controller confirms that he’s got the US Air flight on his radar. That’s a really icky time for things to fall through the cracks.

There’s also a bit of confusion from the departure controller what instructions had been given to the La Crosse-bound aircraft.

Humans make mistakes. That’s going to happen. The FAA will undoubtedly be looking at ways to break the chain of human mistakes that usually are the hallmark of any aviation disaster. Handing out any discipline, hopefully, will be secondary.

Update 5:21 p.m. – Here’s my interview on All Things Considered:

7:50 p.m. – Very important point to consider from Dave Pascoe at

One small thing to keep in mind in your analysis (which after a very quick read I think is right on the mark) – the receivers near KMSP scan between several frequencies. One scans both Tower freqs. The other scans several Approach/Departure freqs. So blocking can and does occur from time to time. It is important to understand that just because a transmission wasn’t heard on the recording doesn’t mean it didn’t happen (like a readback or Tower switching an aircraft to Departure).

I mentioned earlier that there wasn’t a repeat on the tape of the question about the lack of a turn. My guess is that was repeated and wasn’t picked up.

  • Capt. Marcus

    Commercial aircraft and other aircraft are legal to be within 500 feet of each other. A commercial liner would be on an instrument flight plan IFR and another aircraft would be flying under visual rules or VFR at even or odd altitudes plus 500 feet and be quite legal, as they have been for the 50 years I have been flying. The 100 feet near miss means that one of the aircraft was at the wrong altitude or the controller was allowing eather to climb or decend. More facts would be nice.

  • Bob Collins

    Hi, Captain. They were both IFR. At 1100 Z, MSP reported overcast at 900. I don’t have an indication of the tops, however.

  • Sarah

    Thanks for the writeup & audio. It is nice to read an aviation news story where the author understands the story and reports detailed facts rather than uninformed conjecture.

  • Julie

    Not your “typical chain of events”

    There’s something bigger here.


  • SM

    Just to get a bit of perspective here, if I’m understanding the compass headings both planes took off heading northwest, toward downtown MPLS, which would have put the near miss maybe over the intersection of Hwys 62 and 77 or a bit further?

  • Bob Collins
  • Greg C

    The tower can switch you over to departure anywhere from immediately after take off, to several miles out, depending on traffic and coordination with departure.

    It seems that the freight pilot missed the turn out to the south and the tower turned the US Air flight too soon, and before verifying that the 99 was in the turn.

    It pays to keep your ears open when flying out of busy airports like MSP.

    I’ve been flying night freight for more than 20 (accident free) years and spent years flying out of MSP. ATC there is usually top notch but mistakes can happen. The US Air crew did a good job and everybody got lucky this time.

  • Matt Steele

    Sounds like Mr. Beech 99 needs a new radio, by the way.

  • Bob Collins

    The aviation system is full of hard-working people doing the best they can. Sometimes mistakes happen and the system tries as much as possible to isolate those from compounding errors.

    I once landed on a runway in the wrong direction. It was stupid mistake. I guarantee nobody is beating themselves up more than the people involved here and they’ve probably come up with a dozen ways that they’ll personally do things differently to keep it from happening again.

    The best thing aviation has going for it is that (a) it’s a safe system and (b) there’s a personal stake in making it even safer.

  • Fred

    Not to sound goulish,but if they had collided givin the altitude and direction,any guess as to where they would have gone down?

  • Fred

    oops..sorry my last post. comp was on standby and didn’t see your post bob. thanks!

  • Ryan S.

    First off, correct me if I’m wrong, but these are not the “tapes” from ATC. These are the recordings done by LiveATC.Net, these recordings often do not contain the full audio. They are normally scanners that frequency hop between multiple frequencies and can miss things while they are hopping around. For example, there are 2 tower frequencies (one for each parallel runway) at KMSP and LiveATC aggregates them both onto the same feed. It can’t do this by laying them down over the top of one another, it does it by skipping between the two. The result being there could be gaps.

    From the provided audio, I do not hear any acknowledgment from the Bemiji flight of their take-off (T/O) instructions. This would be necessary to confirm whether they acknowledged the take-off clearance AND the left turn to 180.

    Second, and as stated above, the pilot is not required to request a hand-off to departure control. The tower controller is responsible for initiating the handoff.

    Also, please note that these operations were conducted under IFR (Instrument Flight Rules). There are differences between IFR flight operations and VFR (visual flight rules) flight operations. It would be prudent to consult an IFR rated pilot about IFR requirements. In order to operate under IFR as the pilot-in-command, a person must obtain an instrument rating. This is different than simply having a pilot’s license.

  • One note… Cactus isn’t “Air West” it was America West Airlines, which purchased USAirways, but uses the USAirways brand. (Although they’re still not really integrated.)

  • Bob Collins

    America West. Duh, of course. My mistake. Sorry. I also thought US Air bought AW, not the other way around. Thanks for setting me straight.

  • Bob Collins

    I telescoped the recording. I don’t remember whether I included the readback or not. I heard it but I don’t remember if I included it, Ryan. Can you confirm whether pilots are or are not required to readback a clearance to take off w.r.t. the turn?

    // There are differences between IFR flight operations and VFR (visual flight rules) flight operations.

    What do you think might be the differences here?

    One interesting aspect here is when the Airbus inquired about the traffic, departure said he “thought” the plane was heading straight out. What manner of communication is there between tower and departure — if any — so that departure would know traffic isn’t complying with a previous controller’s instructions or even what those previous instructions are?

  • Ryan S.

    Well, I’m not an aviation lawyer… So I can only cite references.

    It is my understanding that the Aeronautical Information Manual (AIM) is not regulatory in nature, however, supplements the Federal Aviation Regulations (FAR).

    Looking in the FAR I don’t see any requirements for read back… FAR 91.123 details that when an ATC clearance has been obtained, no pilot may deviate from it (unless certain conditions occur, ie emergency).

    Moving to the AIM the references I find are in 4-4-7 b that states any pilot of an airborne aircraft should read back such items as assigned altitudes and vectors.

    Then, looking in the Air Traffic Control Order 7110.65 we find pilots are expected to promptly initiate a turn when assigned by ATC.

    As for differences, unlike VFR when separation is the responsibility of the pilot, under IFR certain responsibilities are transferred to the air traffic controller. This can even include pilot hiccups. For instance, the controller could have held the Airbus on runway heading until visually confirming that the Bemiji was turning and that no conflict would occur.

    As far as the departure controller thinking he was issued a straight out departure, he could have been guessing just by the fact that the aircraft wasn’t turning on his scope. I’m not a controller, my understanding is that tower controllers and approach controllers communicate via “hot lines.” It’s my understanding that the Minneapolis approach/departure controllers are not even located on the MSP field (Tower is though).

    It’s my understanding that tower coordinates a release for a flight’s take off clearance with the overlying ATC controller. And that release is predicated upon certain conditions, such as flight path.

  • Bob Collins

    Thanks for that, Ryan.

    // It’s my understanding that the Minneapolis approach/departure controllers are not even located on the MSP field

    I believe they’re in Farmington.

  • Mike E

    It is a bit odd that the beech aircraft was not over to departure already. Normally at MSP, that tends to occur from .5 miles to 2 out.

    Another item is that it sounded like the UA flight was not aware of what the beech aircraft was going to do. at MSP one of two things occur:

    during slow times, one controller handles both runways, so the tower controller would wait to hand over both aircrafts to departure until after they are past eachother, if both are turning the same direction.

    or, if two controllers are running the tower, they will notify both aircrafts that XX will passing behind/in front of you. This is because that they will have one of the aircrafts climb to 8000, and the other stay at a lower altitude, 2000ft to 3000ft until they have crossed. I think in this case the airbus should have, and was climbing.

    Another item is that SOP is to announce your altitude when changing frequencies. The airbus did not when he switched to departure, so we not know how close he was to that 8000ft mark. I think I can understand why that was missed, he was a bit pre-occupied.

  • Skyler B.

    Unless it has changed in the last few years, MSP Approach/Departure Control (radar room- no windows) is located a few floors below the tower cab (the top of the control tower with glass all around).

    Greg C is correct with regard to when a tower controller will “hand” or transfer control of the airplane to Departure Control. Typically, Approach/Departure Control will handle planes usually between 30-40 miles from the main airport (i.e. .MSP), and usually up to around 17,000 feet. Somewhere around the transition altitude or distance, Departure Control will “hand” or transfer control of the airplane to Minneapolis Center (ZMP). They are located in a huge radar room in Farmington, MN.

  • Larry B

    Here are some facts:

    Within the control tower at MSP, one controller works all traffic on runway 30L and another controller works all the traffic on runway 30R. These controllers coordinate between each other when an aircraft will require a turn across the other’s departure cooridor – as happened in this incident.

    These 2 tower controllers each had an aircraft that required turns through each other’s sky.

    Somehow a miscoordination occured or a pilot didn’t properly follow an instruction to turn – ie, possibly BMJ in the BE99, resulting in USAirways running over the BE99 just west northwest of MSP.

    Tower will typically handoff all departures to Minneapolis Departure, located on the first floor – colocated with the tower at MSP, within 1/2 mile of the departure end of the runway. The departure controllers will then climb traffic to as high as 17,000 feet and once clear of all other traffic, hand them off to Minneapolis Center, located in Farmington, MN.

  • RJ

    Thankfully the Catcus flight had TCAS, question is why did the BMJ46 not have it?

  • Bob Collins

    They’re not required to at the moment. There are some cheap alternatives that supply some traffic information that you can install in a plane for as little as $1200 now (I’m putting one in the plane I’m building).

    It may become somewhat academic in a few years. The country is moving toward a system called ADS-B, which determines a plane’s position by a GPS, relays it to ground stations AND other aircraft. It’s part of the so-called NEXTGEN in air traffic control.

    As I understand it, ADS-B will require even private pilots to install ADS-B compatible transceivers that at minimum are expected to cost about $5,000. As you might expect, some pilots aren’t that happy about the idea. Some are.

  • Paul

    Worried passenger – I noticed the last time (13 Oct) when I left MSP from 30R they had us take a left turn right across 30L departure for an east bound heading. From 30R it is usually always a right turn for an east heading and climb out. I agree with Juliee I think there is something more to this. I hope no one is trying to stack departures in the turn.

  • Larry B

    Paul, if the airline you flew filed a route over Dells (DLL) you’ll get that turn to the north, but it goes through a very crowded arrival route to Minneapolis from the northeast. Most airlines file over Nodine (ODI) to avoid that, and thus get that left turn around the south of Minneapolis around the farmington area.

    Everything leaving the twin cities departs in a long line – there’s no such thing as stacking departures anywhere. They are vectored by the departure controller and basiclly fanned out on various headings depending on their destinations… typically leaving the area on 7 different courses.

    Regarding TCAS, there is no requirement for these small BMJ cargo aircraft to have it. These planes are extremely dated, many from the 1960’s.

  • Rick T

    This type of departure procedure is very common at busy airports. I believe the Tower Controller kept the BE-99 on his frequency to ensure the turn. Maybe not, but it would not be uncommon practice, he had issued an initial turn to the BE-99. As has been said; one Controller works one runway and another Controller works the other runway. Coordination is probably required, however, I would think that they (FAA), has a “land-line” recording that explains what coordination took place. We may not have access to that information until an official report is made available to the public. I’m sure some parts of this practice have been amended or changed completely

  • Bob A.

    Having been a tower and approach controller for 28 years I think it is common practice to clear two departures off of parallel runways, as long as there is some kind of insured separation. It may be published or issued by the controller. I heard a two part clearance, turn left and cleared for take-off. No read-back from the Beech about the turn. The Beech goes into the clouds, so the only bailout is the radar monitor in the tower, not a frequency change to departure. It’s a tower function to insure initial departure separation, coordination with departure about you not doing your job causes problems and wastes time. The next problem is the heading for the jet. The tower and approach control have a letter of agreement about departure headings. The controller issued the left turn to the jet from rote memory, that ‘s what the letter says. Bottom line; ATC issues instructions and insures compliance either through read-backs, or observed turns visually or on the radar. MSP might look into a monitor on the local controller position, because without TCAS the jet had no clue where the other was.