Mid-Air Collision over the Potomac
A tragic mid-air collision took place near Washington DC on the 29th of January 2025. A Bombardier CRJ700, operating as American Eagle flight 5342 and a Sikorsky H-60 “Black Hawk” military helicopter crashed into each other at around 200 to 300 feet above ground level over the Potomac River.
Flight 5342 was a scheduled domestic passenger flight from Wichita, Kansas on final approach for runway 33 at Ronald Reagan Washington National Airport (DCA). , The CRJ700, registered in the US as N709PS, carried sixty passengers and four crew
According to FlightRadar24, the last data received from the aircraft was for level flight at 400 feet.
The weather at the time was with a westerly wind with ten miles visibility.
The US Army helicopter, registered as 00-26860, departed from Davison Army Airfield on a training mission. There were three on board, a captain, a chief warrant officer 2 and a staff sergeant. They were were flying south along the Potomac River as a part of an annual proficiency flight and night evaluation. The maximum allowed altitude for the route that the helicopter was travelling, transiting from route 1 to route 4, was 200 feet above ground level.
The collision took place at 20:48 local time (01:48 UTC) and has been reported at less than 200 feet altitude. Over three hundred responders were deployed, fighting against the icy conditions. There were no survivors.
Audio Player
The above ATCLive recording is a merge of two frequencies so that the aircraft and the helicopter interactions are audible at the same time. You can also listen to the two separate frequencies on VASAviation. The New York Times report that the two controller roles are usually merged at 21:30 but were merged earlier that day, with one controller managing both frequencies, as the controller needed to leave early.
The initial back and forth the Tower controller and flight 5342 were standard for the approach. American 3130 and American 427 were also on approach behind the accident aircraft.
The relevant interactions start with the controller warning the helicopter about the descending flight.
Tower Controller: PAT25, traffic just south of the Woodrow Bridge, a CRJ at 1,200 feet setting up for runway 33.
Helicopter: PAT25 has the traffic in sight. Request visual separation.
Tower Controller: Visual separation approved.
Tower Controller: PAT25, do you have the CRJ in sight?
Tower Controller: PAT25, pass behind the CRJ.
Helicopter: PAT25 has the aircraft in sight, request visual separation.
Tower Controller: Vis[ual] sep[aration] approved.
Unidentified Pilot in the air: Tower did you just see that?
Tower Controller: American 427, American 3130, go around. Turn left, heading 350, climb and maintain 3,000 foot.
It is yet unclear exactly how they ended up in conflict. The helicopter crew’s request for visual separation shifts responsibility to the helicopter crew to see and avoid the aircraft. The controller gave a specific instruction to the helicopter to pass behind the aircraft after agreeing to visual separation. The helicopter crew’s second request for visual separation may indicate that the helicopter did not intend to pass behind the landing aircraft, as that request was not acknowledged. They obviously believed that they had the CRJ700 in sight but with three aircraft inbound to DCA, a key question will be which aircraft did the helicopter crew actually have in sight.
This Google Earth image posted without attribution on the New York Times claims to show the flight path of the CRJ700 passenger jet, coming in from the top right, and the helicopter, travelling left to right.
This video by Captain Steeeve breaks down the ATC communications; I’ve jumped ahead to the Ground Frequency transmissions after the crash which I found particularly interesting.
Note that the helicopter did not necessarily have a Traffic Control Avoidance System (TCAS) installed, while the CRJ700’s TCAS would have been inhibited as it descended through 1,000 feet above the ground 400 feet above the ground (see comment below from Philippe).
. This is to avoid constant warnings with aircraft nearby and on the ground, which in heavily trafficked areas quickly turn TCAS warnings into a distracting nuisance. In addition, TCAS offers resolution advisories meant to be followed at altitude and could lead to an aircraft descending too close to the ground. Air Traffic Control and visual separation are generally more reliable at low altitudes.
This is Earthcam footage looking southwest, showing the impact:
The NTSB’s media briefing confirmed that taking part in the investigation are:
- TSB Canada
- MHIRG (previously known as Bombadier)
- PSA Airlines
- GE Aerospace
- Sikorksy
- FAA
- NATCA (representing air traffic controllers)
- ALPA (representing pilots)
- AFA (representing cabin crew)
- US Army
Since then, the NTSB has confirmed that the flight data recorder and the cockpit voice recorder from the CRJ700 have been recovered. The recovery teams are still searching for the helicopter’s recorder.
Most of the rest of the briefing was spent asking press not to speculate on causes and that the NTSB will of course be considering all aspects including human factors.
Despite this careful approach, US President Donald Trump speculated that the crash was somehow caused by DEI (diversity, equity, inclusion) hiring, implying fault with the air traffic controller. However, when asked whether there was any evidence to support this claim, he said “It just could have been.” He backed this assertion with “I have common sense” and claimed that the Obama administration thought that personnel were “too white”. The Washington Post published a fact check which refutes Trump’s claims and details the actual FAA hiring practices under Obama, Trump and Biden.
The New York Times reported that the tower at Reagan airport has been understaffed for years and that an internal FAA report said that the staffing was “not normal for the time of day and volume of traffic”. On the ATCLive recording, it’s clear that a single controller was dealing with arriving and departing aircraft as well as local helicopters on a different frequency. According to the New York Times, one controller left early and the supervisor combined the two roles, leaving the one controller to deal with the evening traffic.
This crash ends a sixteen-year streak without a fatal US commercial airline crash. The previous fatal crash was Colgan Air flight 3407 in 2009. Key factors in this investigation will likely include the responsibility for separation and conflicting expectations of flight paths. The NTSB have confirmed that they will release a preliminary report in thirty days. Their investigation will be crucial in determining the exact sequence of events and digging past the obvious to identify the underlying causes.
It will be interesting to see what exactly went wrong regarding TCAS. Presumably, it might be related to how the helicopter handled their transponder, being a military aircraft. TCAS is not actually disabled below 1000′, it just doesn’t create resolution advisories (“TA only”). It should, however, warn of proximity intruder traffic via a traffic advisory (TA). TAs would also work if the helicopter itself is not equipped with TCAS as long as it had a transponder transmitting in mode A (even mode C, altitude transmissions, would not be required for a TA, although in this case the target shown to the other aircraft would not include any altitude information).
Two things that I found totally surprising
The heavily trafficked helicopter corridor directly intersects the approach end of runway 33 at low altitude when the pilots are heavily occupied with landing or takeoff.
In such crowded airspace, traffic control allowed for visual “see and avoid“ at night in the one of the busiest air corridors on the East Coast. I thought this practice ended decades ago in near proximity to major airports.
I’m typed in the CRJ and I’ll have to correct myself regarding my last comment, as I’ve just read it up in the OM-B: While only RAs are inhibited below 1000′, TAs are inhibited below 400′ in descent (and 600′ in climb). Additionally, TAs are not created for targets below 380′.
The crew should have received a TA regarding the first inhibition (the A/C would have been above 400′ when the helicopter became an intruder), but as the collision apparently occured at 200′, the helicopter would have been to low for a TA. Really a tragic event.
Thanks for this! I’ve updated the above to reference y our comment.
Anecdotally, I can’t find any such restriction in the manuals of the A320, that I used to fly before. While the Airbus also inhibits RAs below 1000′, it doesn’t seem to have a restriction on intruder altitude. In this case, that may have been crucial…
With the number of near misses that have increased over the last few years, it was not a matter of if, but when. Understaffed ATCs, hardware that is well behind the state of the art, and a federal government that has continuously cut funding-all factors. Plus having military training flights going through approach paths increases the risk of contact. Regardless, it’s tragic and my heart goes out to the families and friends of the victims.
PS: The decision to politicize a tragedy only hours after it happened, instead of offering support and sympathy, was a huge mistake, but one that seems to play into the current political climate. And we are lesser people for allowing it.
Looking at one video of the incident, there is another aircraft climbing away above the collision. I just wonder if the helicopter pilot saw this aircraft when he responded to having visual? Thinking he was in the clear and not seeing the aircraft he was warned about? Pure speculation on
my part. As Arthur commented, I too wonder why helicopters especially on training flights are allowed to fly directly across the landing flight path
of a busy commercial airport.
As for Trumps suggestions I am speechless!
I just want to say thanks for a clear-eyed presentation of the facts and some really good reference material.
One of the reports I’ve seen in the NYTimes suggests the helicopter wasn’t as low as it was supposed to be (200 feet); another notes that the airplane was originally supposed to land on 36 instead of 33. (36 is by far the longest runway, so it would be preferred for takeoffs.) I haven’t seen any report about possible confusion with other inbound aircraft, but Sylvia’s point is plausible given that ATC’s go-around direction suggests there were two other aircraft on close final; it’s possible the copter pilot didn’t realize that the airplane’s landing lights weren’t pointed at him and thought the next plane was the one to dodge.
But past all this, I’m baffled by anyone anywhere thinking that this was a good place to conduct a refresher/currency exercise; the practice should be done with cones or other subtle markers somewhere out in the boondocks. (If they need to practice following a ragged river, the Potomac much further upstream would be a starting place.) IMO there’s no need be familiar with that piece of riverbank (enough to follow the wild course shown on another graphic); if Baby Trump needed to go somewhere in a hurry he’d ride a Marine cargo-ish helicopter, not a fighter, directly away from DCA toward Andrews AFB. (Yes, Camp David is NW of DC; a copter flight there should still climb eastward from the White House before turning on course.) The new defense secretary will be utterly useless in dealing with this; he’s a talk-show host with no serious command experience, and even the line commanders who haven’t already been driven out know this.
I’ve never piloted at DCA, but I’ve been a passenger there a lot (family, conferences, contract work); riding in and out, and watching from the many nearby parks as planes flew in and out, were always … interesting. Running a nighttime helicopter flight right past there was looking for trouble.
It’s reported that the helicopter belonged to a unit that handled VIP transport and “continuity of government” missions (i.e., evacuation/relocation of government personnel to alternative sites during an emergency in Washington):
https://www.twz.com/air/mid-air-collision-shines-light-on-doomsday-plans-for-nations-capital
Secretary of Defense Pete Hegseth is quoted as saying that the helicopter crew were undertaking “a routine annual retraining of night flights on a standard corridor for continuity of government mission”, so it sounds like it was a refresher/concurrency exercise for a specific route through Washington that they would use on operational missions. That’s not something that can easily be replicated elsewhere.
But I accept your point that it is a hazardous location to be doing any kind of exercise. It wouldn’t surprise me in the least if the investigation discovers that there was a risk assessment done years ago which gradually became more and more out of date as traffic levels increased, but which was still relied on long after it should have been discarded.
https://www.nytimes.com/interactive/2025/01/30/us/washington-dc-plane-crash-helicopter-maps-photos.html shows a strangely jagged route for a helicopter practicing a get-out-of-Dodge maneuver, or even a pick-up-and-run maneuver; that route doesn’t go anywhere near where the story says “key people” would be picked up — in fact, the path veers AWAY from the Pentagon. (Not directly visible in those pictures, but it’s immediately WSW of the lagoon under the overlaid word “Army” in the first picture.)
I suspect you’re right about safety creep; members of Congress keep trying to add flights so they don’t have the modest inconvenience of having to go out to Dulles even for flights to the other coast, per https://en.wikipedia.org/wiki/Ronald_Reagan_Washington_National_Airport#Operations (Driving time from downtown to IAD is variable, but the subway goes all the way out there.) IIUC, this particular flight was added just recently; I saw a quote from a Kansas congressman expressing happiness over it.
I suspect that one way to reduce the likelihood of a collision would have been to fly the mission some time after midnight — but that might have annoyed a lot of well-off people. (The riverfront became highly sought after once the factory/freight/fishing operations were moved away; the poor parts of DC are mostly above and to the left of the photo, although parts of the left side aren’t in great shape either.)
I’m just hoping Trump’s baseless incompetent accusations won’t result in his followers grabbing the pitchforks and going after the ATC controller in question.
It’s so upsetting. That and his “do you want me to swim to the site” comment. It’s so horrible. I came straight here to see Sylvia’s take because I knew she would handle it with dignity.
Thank you Sylvia, for doing what you do.
The helicopter did have a Mode S transponder, and does show up on the flight tracker websites as MLAT target, i.e. the ground stations determine its position, the helicopter does not broadcast it as it would with ADS-B. This also means any kind of “jagged” path is a data artifact, the heli crew actually flew a smooth curve. Regardless of the “training” label, the pilots were qualified and experienced.
I believe that the helicopter would have been indicated on the CRJ’s TCAS screen, but with the runway change and the general workload of a night landing, the crew had other places to direct their attention to.
The runway change itself was not unusual or unsafe per se; switching the CRJ from 36 over to 33 meant the next aircraft could land on 36 without waiting for the CRJ to leave the runway. The traffic controller briefed the helicopter crew that the CRJ was inbound to 33, so they should have expected the CRJ to cross in front of them, but perhaps did not realize this.
What struck me is that the helicopter crew does not read back anything that the controller said to them (and the controller seems way too busy to ask for a readback); is that normal under the circumstances, i.e. when they’re not giving direct instructions?
Conducting night visual operations in an urban space with dense air traffic seems like a recipe for failure, with not a lot of margin for error.
I’m not going to comment on the President that the American people have chosen to elect; but the fact that they did so twice speaks volumes.
I don’t know what the helicopter pilots were taught about readback in civilian airspace, but the controller being too busy to ask for a readback is presumably an effect of one controller having been allowed to leave early (as noted by Sylvia and in various news reports), so the remaining controller was doing two jobs. ISTM this is bad management, but I haven’t seen why the leave was allowed or whether the supervisor knew when he approved the leave that there was going to be a helicopter in the neighborhood.
It’s doubtful that the supervisor knew that there would be a training flight that night; the military doesn’t normally file flight plans with the FAA
Normally, no. But do they not file a flight plan when they’re going to be in very crowded airspace? Can they just show up and say “Hi, we want to cross one of your approach paths.”? I get that they might want not to file a plan in a real emergency — but in a real emergency National would probably be shut down, just as it was on 9/11.
It’s drilled into every pilot’s head to read back instructions to ensure that everyone is on the same page. Even military pilots do it, and why the helicopter pilot didn’t will I think be a contributing factor.
Not that it was ever a good time to be a minority in the US but it’s particularly bad and scary now.
Mendel: why would diverting to a different runway make room for the plane behind? Isn’t there supposed to be enough separation in the air to allow for runway clearance? If anything, I’d think the diversion would increase the hazard, because the following plane would be landing (or at least trying to land) on a crossing runway (1 vs 33).
Because the CRJ was landing on a cross runway, and would only momentarily cross the north south runway, the distance between aircraft on final could be decreased, thereby allowing more operations in the airport. That’s the reason they do this.
Scenario 1: Aircraft 1 (A1) lands on 36. It needs to slow down at the end of the runway, and then find a taxiway to get off the runway, before it is safe for aircraft 2 (A2) to land.
Scenario 2: Aircraft 1 lands on 33. Once it is past the intersection with 36, it’s safe for aircraft 2 to land on 36. The rollout etc. of A1 does not affect A2.
In scenario 2, the aircraft can be spaced closer and still be safe from each other.
so….
1) obviously it’s runway 01, not 36, I don’t know where I got that from
2) from comments on avherald, it seems that the sidestep to 33 also served as a means to increase separation with a preceding departure: the CRJ touches down later, and the preceding aircraft only needs to be past the intersection to be safe.
We’ve also had some near-miss events where this kind of procedure is cutting it close, but it’s still preferable to using a single runway only.
It was runway 36 when I was flying to/from DCA regularly, decade ago; https://en.wikipedia.org/wiki/Ronald_Reagan_Washington_National_Airport#Expansion_and_restrictions says that it and the \other/ shorter runway (formerly 03/21, now 04/22) were renumbered in 1999 due to magnetic north drifting (as it is always doing). 33 wasn’t presumably because its direction was sufficiently counterclockwise that it didn’t go over 335 magnetic at that time, but its time may come.
I’ll gladly speak to it Mendel. A lot of my fellow Americans are easily duped and Trump told them what they wanted to hear to be elected. A lot of the more intelligent Americans I know were in willful ignorance, “No one is dumb enough to elect a convicted felon”, so they just didn’t vote.
And, well……..As an American that’s traveled the globe I already know how other countries viewed us. Now they just see us as racist and pathetic. Which, to be fair, I live in the south. I’ll leave it there.
Aaaanndd, the Trump-led rumors have cascaded; a Navy pilot who is trans was claimed to be responsible for the crash, even though she’s alive: https://www.nytimes.com/2025/01/31/us/virginia-pilot-black-hawk-helicopter-crash.html
Because it fits his narrative that DEI=America’s problems, and that transgender people are inherently bad.
What the hell is wrong with people? Good grief.
To be clear, my “what the hell is wrong with people” is meant for people believeing that a trans Navy pilot who is still alive is repsonsible for the crash somehow, not the DEI comment.
The whole DEI rehetoric is just coded language for “it’s okay to discriminate against anybody who’s not a cis-vanilla person who’s orginially from Europe”, IMHO.
A commentator here in Australia described it as “Make America Hate Again”, which hits the nail on the head IMO.
Sad for us; a strong US was good for us.
A commentator here in Australia described it as “Make America Hate Again”, which hits the nail on the head IMO.
Sad for us; a strong US was good for us.
I think it was Juan Brown who posted a Youtube video showing ADS-B for the helicopter before impact showing 350ft ALT. Immediately prior to contact AA 5342 showed an ALT of 375ft, yet the NYT article which you linked claims 5342 was down to 275ft. I am more inclined to believe impact occurred at 375 ft . Can anybody clear up conflicting information?
It seems the Blackhawk was well above their 200ft ceiling.
The initial article was probably working with what they thought they had when it was written. The more recent article https://www.nytimes.com/2025/02/01/us/politics/american-airlines-crash-ntsb.html says that the NTSB investigator reports the plane moved slightly upward, possibly attempting to dodge, and places the collision at 300-350 feet.
The NTSB is going to clear this up. From their 3rd briefing (via AVHerald):
There’s enough information, this will resolve itself if you sit back and wait.
The helicopter route can be used by any helicopter.
During the day, there are two controllers working, so it makes sense to have two separate frequencies.
I get it why the military use discreet military radio frequencies, but they may as well have been speaking French.
If this was not a VIP flight Everybody should have been using the same frequency, that way pilots have a fighting chance to avoid collision when the controller is too busy..
If it were up to me, all helicopter routes would be forced out 4 nautical miles away from the airport and a new specific joining procedure created for helicopters coming to the airport itself.
Poor kids on that flight were sacrificed for the convenience of VIP helicopters.
Per one report, the military helicopters, as well as civilian helicopters, frequently interact with Reagan ATC on a different frequency than airline traffic.
Not sure of the reason. Military traffic would also be in contact with their command on secure frequencies, but ATC comms use normal civilian channels and are under the same ATC control.
This is because of several fatal military plane collisions with airliners in the 1950’s and 1960’s, after which the military and civilian traffic control was unified.
There was a 1958 United Air Lines Flight 736, a Douglas DC-7 which collided with an Air Force F-100F fighter jet at 21,000 feet over Las Vegas. This and a smaller collision in May 1958 were the final nails in the coffin for the CAA, which was doing an inadequate job on safety. The CAA was dissolved and the FAA created.
Ref: https://news3lv.com/news/local/1958-las-vegas-military-jet-and-passenger-plane-collision-sparks-the-formation-of-the-faa
Las Vegas
I think you mean the CAB, not the CAA?
I am trying to argue for common sense. In proximity to a busy airport everybody should be on the same page.
Not only was the Helo on a UHF frequency, but the ADS-B -out was switched off too! This just is not acceptable. 67 people who never signed up to the military to take such risks had their lives tossed aside for a reckless training flight.
The Civil Aeronautics Board (CAB) regulated the airline market and preceded the NTSB with respect to safety.
The frequencies differ for rotorcraft and fixed wing aircraft because they each have different needs and are routed differently, usually by two different controllers.
It is far too early to slander the helicopter crew as “reckless”.