top of page

Blog

After the accident, it was immediately clear that helicopter traffic operating on Helicopter Route 4 posed a risk of a midair collision with aircraft landing on runway 33 at Ronald Reagan National Airport (DCA). The FAA temporarily shut-down helicopter traffic on Route 4 and this prohibition was made permanent after the NTSB issued its urgent safety recommendation A-25-1 on 7 March 2025.


Applying the "Swiss cheese" model to this tragic accident, the following circumstances contributed to the midair collision between a CRJ airliner and a military helicopter:


  • At 20:33, the helicopter crew requested Helicopter Route 1 to Route 4 , which the tower controller approved.

  • At 20:39 the approach controller cleared the airliner for a straight-in visual approach to runway 1.

  • At 20:43, the tower controller asked if the airliner could switch to runway 33. After a brief discussion, the crew agreed to this request and the tower controller cleared the airliner to land on runway 33.

    • Note 1: This required a change to the previously briefed plan for approach and landing on runway 1, and a more demanding nighttime, circling, visual approach to runway 33.

  • At 20:46, the tower controller informed the helicopter that there was traffic just south of the Wilson Bridge, a CRJ at 1200 ft circling to runway 33. Although at this time the helicopter was about 6.5 nautical miles (12 km) north of the airliner's position and travelling west, the crew reported that they had the traffic in sight and requested to maintain visual separation. The tower controller approved visual separation.

    • Note 2: Upon accepting visual separation, a pilot assumes responsibility for avoiding the other aircraft, based on the principle of "see and avoid"; minimum horizontal and vertical safe separation distances no longer apply.

    • Note 3: The tower controller's transmission was garbled and the word "circling" was not registered on the helicopter's CVR.

    • Note 4: The tower controller subsequently cleared two more airliners to land on runway 1 and the landing lights of these aircraft would have been visible to the crew of the helicopter as it turned to fly south on Route 4.

  • At 20:47:39, 20 seconds before impact, the tower controller asked the helicopter if it had the CRJ in sight. A conflict alert in the control tower was audible in this transmission.

    • Note 5: This call and the alarm means that the tower controller was now aware that the two aircraft were on converging courses, but he was still relying on the helicopter to maintain "visual separation". He was probably also relying on an assumed vertical separation to avoid a collision, despite the fact that the radar screen only gives the aircraft's height in hundreds of feet.

  • At 20:47.40, 19 seconds before impact, the airliner's Traffic alert and Collision Avoidance System (TCAS) issued an audible Traffic Advisory (TA): "Traffic, traffic".

    • Note 6: A TCAS Traffic Advisory (TA) is designed to activate 20-48 seconds before a potential collision: the crew is required to attempt visual contact and to be prepared to manoeuvre if a Resolution Advisory (RA) occurs. At this time, the helicopter was about 0.95 nautical miles (1.8 km) directly ahead and slightly below the airliner.

    • Note 7: A Resolution Advisory (RA) is designed to activate 15-35 seconds before a potential collision and issues a command to "Climb, climb" or to "Descend, descend"; however, RA's are inhibited below 1000 ft radio altimeter height.

  • At 20:47:42, 17 seconds before impact, the tower controller instructed the helicopter to pass behind the CRJ.

    • Note 8: The phrase "pass behind the" was not registered on the helicopter's CVR because of a simultaneous transmission by the helicopter's radio.

  • At 20:47:44, 15 seconds before impact, the helicopter indicated that the traffic was in sight and again requested visual separation which was again approved by the tower controller.

  • At 20:47:58, one second before impact, the airliner's elevators were deflected to near their maximum nose-up travel and the airplane reached 9° of nose-up pitch.

    • Note 9: This strongly implies that, finally but too late, the airliner's crew were aware of the impending danger, either because they saw the helicopter or responded to the TCAS alert and the graphic indication of a target on its screen.


As highlighted in the notes above, the different circumstances which "lined up the holes in the cheese", were as follows:

  1. The tower controller´s request to the airliner to change its approach from runway 1 to runway 33, when he had previously cleared the helicopter to fly Route 4.

  2. The airliner crew's acceptance of a nighttime, circling, visual approach to runway 33. The crew were not obliged to accept and had several valid reasons to reject this request.

  3. The tower controller identified a potential loss of minimum separation between the airliner and the helicopter, which in VFR is 1,5 miles horizontal and 500 ft vertical. He gave the helicopter crew the type, location, altitude and intended path of the airliner. This induced the helicopter crew to request maintaining visual separation, stating that it had the airliner in sight. However, at night in an urban environment, at a distance of 12 km and with other aircraft heading directly towards you with their landing lights on, it is unlikely that you would correctly identify an aircraft that is now heading away from you on a circling approach pattern. This is even without considering that the crew of the helicopter were wearing night vision goggles, designed to highlight unlit or poorly lit obstacles but with a very limited field of view. The conclusion is that the helicopter crew never correctly identified the CRJ circling to land on runway 33 and in fact probably never saw the airliner because they were focusing on another aircraft.

  4. The tower controller missed a last opportunity to warn the crew of the airliner when the collision alert sounded in the tower cab; instead, he again asked the helicopter if he had the CRJ in sight.

  5. When the TCAS alert sounded in the airliner's cockpit, the crew should have been able to observe a target at the 12 o´clock position on the TCAS screen, they should have been on the lookout outside and they should have been able to see the helicopter's position and anticollision lights. That they did not see the helicopter in sufficient time to avoid the collision is just another example of the failure of the "see and avoid" concept.


The final "hole in the cheese" was the helicopter's deviation from the maximum height of 200 ft stipulated for this sector of Route 4. According to the NTSB Preliminary Report, at the time of the collision the helicopter was at a radio altitude of 278 ft but it is not clear if the helicopter had a radio altimeter display in the cockpit. The crew would have been relying on the barometric altimeters and about four minutes before the crash the CVR recorded the pilot saying that they were at 300 ft while the instructor pilot (IP) said that they were at 400 ft. Neither pilot commented on this 100 ft discrepancy and this will obviously be part of the continuing investigation.


Notes 3 and 8 highlight another serious flaw in the present ATC system, but I will cover this aspect in my next post.

 
 
  • Writer: Rodriag Symington
    Rodriag Symington
  • Mar 25
  • 2 min read

I will now restate those ATC practices which I consider unsafe and what should be done about it.

1.        “See and avoid” and “visual separation” applied to airline traffic. These principles are clearly unsafe and should not be relied upon to maintain safe separation between airliners and other traffic in controlled airspace. It should only be applied between airliners in daylight and good visibility when there is no possibility of misidentification, for example in the case of airliners following each other to land or when landing on parallel runways.

2.      The US practice of clearing multiple aircraft to land pending the runway to be clear of traffic. This “US only” practice does not expedite traffic but allows the controller to “take his eyes off the ball” at a critical moment and has led to several serious incidents at US airports. A go-around at the last moment presents an increased risk and should be avoided. This is an unsafe practice which should be discontinued.

3.      Nighttime visual approaches. These entail greater risks than daytime visual approaches and for this reason are prohibited by several airlines. This is an unsafe practice for airliners and should be restricted to general aviation or when the runway in use does not have an instrument landing system. Naturally, this prohibition will reduce capacity at some major airports, especially in the northern winter months, but this is a necessary sacrifice in favor of safety.

4.      Taxi clearances which include permission to enter or to cross a runway without stopping. The danger of a clearance which includes permission to enter or to cross a runway is that it is open to misinterpretation or that the crew can mistake one runway for another or for a taxiway, which is not uncommon. A safer practice is to rule that all aircraft should always stop before entering or crossing any runway; at controlled airports a positive clearance from ATC would be required before proceeding to enter or cross a runway.

Of course, I do not pretend that these ingrained practices will be changed any time soon, much less just because I consider that they are unsafe. An adequate procedure would be for the NTSB to study them and make recommendations to the FAA. The FAA would then convene a special committee, which would ideally include pilots as well as ATC specialists, to consider if any changes should be implemented.

 
 
  • Writer: Rodriag Symington
    Rodriag Symington
  • Mar 25
  • 3 min read

Continuing with my call for a reappraisal of potentially unsafe practices in the US air traffic control system, I will now add the following to my list:


“Hold short of Runway X”. Evidently, this instruction, which requires a correct readback, was instituted in order to enhance safety by preventing an aircraft from entering an active runway. However, I believe that in fact it is part of an unsafe practice that allows ATC to clear an aircraft to cross one or more runways before coming to a stop. A safer practice is to rule that all aircraft should always stop before entering or crossing any runway; at controlled airports a positive clearance from ATC would be required before proceeding to enter or cross a runway.


The danger of a clearance which includes permission to enter or to cross a runway is that it is open to misinterpretation or that the crew can mistake one runway for another or for a taxiway, which is not uncommon.


For example, in the ground collision between a Japan Air Lines Airbus A350 and a Japan Coast Guard DHC-8 which occurred at Tokyo Haneda Airport on 2nd January 2024, ATC instructed the DHC-8 to “taxi to holding point C5”, but the surviving captain of this aircraft asserted just after the accident that he had been cleared for take-off. The DHC-8 proceeded to enter and line-up on the runway, where it was hit by the landing A350. Of the six crew on-board the DHC-8, only the captain survived; all 379 occupants of the A350 managed to evacuate the aircraft before it was completely destroyed by fire.


Another recent example was the near-miss at Chicago Midway International Airport (MDW) on 25th February 2025. A Bombardier Challenger 350 CL-35 was operating under General Aircraft (GA) rules as Flexjet flight 560 and was initially given a clearance to taxi to runway 22L via taxiway F and to hold short of runway 4L. When they were on taxiway F, ATC instructed them to turn left onto runway 04L, cross runway 31L and then hold short of runway 31C. The flight crew initially read back the instructions incorrectly; the ground controller immediately reissued the instructions and received a correct readback. After turning left onto the runway 04L with a width of 150 ft, the crew crossed runway 31L, mistaking it for a taxiway and then proceeded to cross runway 31C believing it was runway 31L.[1] At that moment a Boeing 737-800 operating as Southwest Airlines flight 2504 with 136 passengers and six crew, was about to land on runway 31C. The airliner was below 100 ft when the co-pilot realized that the business jet was going to cross runway 31C; he called for a go-around and the captain executed it, with the airliner reaching within 25 ft of the runway surface before climbing away. The Boeing passed with a vertical separation of 75 ft and a horizontal separation of 200 ft behind the Bombardier. In this case, the change I propose would have required the business jet to stop before entering runway 04L and before crossing runways 31L and 31C,[2] but better safe than sorry. In 2001, another business jet strayed onto the runway at Milan-Linate Airport in Italy, killing all 110 passengers and crew on an SAS airliner, four more persons on the ground as well as the four occupants of the smaller aircraft.  


[1] It is noted that runway 31L/13R has a width of just 58 ft and its runway signs and hold-short line on runway 04L can be difficult to see from an aircraft turning onto runway 04L from taxiway F because these markings are just 75 ft from the centerline of taxiway F and just where the nosewheel of the aircraft would be straightening out. Runway 31L/13R has no surface markings crossing runway 04L/22R. From the lack of tire marks on runways 13L/31R and 04L/22R, it appears that these are mostly used as taxiways.

[2] It appears that runway 31R/13L was decommissioned between 2023 and 2024; it is now a taxiway.

 
 
  • Facebook
  • Twitter
  • Instagram

Aviation Insights

© 2025 by Aviation Insights.
All rights reserved.

Contact

Ask us anything

Please add any comments.

bottom of page