I remember learning during my pilot training that when you were operating an aircraft at a towered airport, you would be talking to a tower controller when you were landing on a runway or getting cleared to take off from one. I was also told that when you taxied to a runway for takeoff, taxied off the runway, or taxied to a ramp area to park, you would be talking to a different ground controller on a separate frequency.
We still have separate ground and tower frequencies at most controlled airports (Class D, C, and B), but often the same person manages traffic in both ground operations and the runway area. They operate on “combined” ground and tower operations, which increases workload for those controllers, requires spreading attention over larger areas of the airport, and increases the potential for things to get missed and for bad things to happen.
This week’s accident at LaGuardia exemplifies this problem. I know that the NTSB and FAA investigations into this tragic event will yield many lessons. In the meantime, I have a proposal to add another layer of safety to the system as a check against events like this.
Ground controllers should be limited to controlling ground operations, not operations on runways, and do so on the ground control frequency. Tower controllers should only control what happens on the runways and do so on the
tower frequency.
What do I mean by this?
I mean that we modify our procedures at towered airports so that a ground controller is in place to manage aircraft, vehicles, and pedestrians navigating the airport’s ground surfaces. The ground controller would be able to control things moving around the airport, but would never be allowed to give a clearance for aircraft, firetrucks, plows, whatever it happens to be, to cross a runway.
That is what the tower controller, on the tower frequency, is there for.
How would this look in operations?
Simple. A pilot, a fire truck driver, or an aircraft operations vehicle might call the ground controller from a ramp to request taxiing to a runway or another area of the airport.
That ground controller would issue a clearance for taxi, but would have to issue a “hold short” at the point at which that vehicle would encounter a runway. It might sound like this, “Operations vehicle 21 cleared to taxi Alpha, Bravo, Charlie, hold short runway Nine Left.”
This would be the limit of the taxi clearance, but it would end with a couple of extra words, “contact tower 119.0” (or whatever applicable tower frequency was appropriate at the airport in question.
When reaching this point, the vehicle or aircraft would switch from the ground control frequency and contact the tower controller who had control over the active runway. The tower controller, solely responsible for what is happening on that runway, would issue a clearance or a continue-to-hold-short clearance. They, and they alone, would be the ones to allow an aircraft or vehicle to cross a runway on which takeoffs or landings might be in progress or imminent.
Once the tower allowed the aircraft or vehicle to cross the runway, when it was clear, they might offer a clearance that sounded like, “Aircraft 123, cross runway nine left, contact ground on the other side.”
Upon clearing the runway crossing, the aircraft or vehicle would again contact ground control to continue taxi. If they encountered any runways again, the process might repeat. The point of modifying the process is to ensure that a tower controller is not distracted by other aircraft or vehicle movement from what is happening on the runway, where high-speed collisions can tragically occur.
It keeps focus on the key areas of concern, where things are moving quickly, and would inhibit a ground controller from allowing an aircraft to cross an active runway. Some will argue that at many airports, tower controllers operate both ground and tower frequencies, controlling both areas. That staffing levels “can allow” this and that we are critically low staffed in ATC, we “need” to do this.
I understand. And I get it. But this is the United States. We can do better. It will take a bold decision to make a change like this happen. We have to choose to train and fund enough ATC positions to make this happen. In a perfect ATC world, we would do this at every controlled airport. At a minimum, we should start by targeting this at every Class B airport in the United States airspace system. These are the busiest, so we can allocate our staffing and training requirements to ensure enough staff are assigned to them and implement this recommendation.
From here, we should target other controlled airports, including those operated by contract tower operations. Just two days ago, I was at a Class D airport that was running parallel runway operations with one controller operating both ground and tower frequencies. As he did, commendably, a very great job of it, it was a heavy workload. At one point, I counted 16 aircraft in the air in the pattern, inbound, or outbound into the airspace, and trying to taxi around the airport to depart or get to parking.
Splitting that workload between a tower controller and ground controller is how our system was designed to work. We need to get back to doing it instead of combined frequency, which overtaxes the responsibility level and pushes our ATC controllers to the limit of operations.
This isn’t a solution to all potential runway incursions. But it is another step we can take to further limit the potential for large, deadly ones to happen. Each step we take to improve communications and coordination where high traffic volumes of aircraft, vehicles, or people come together in our airspace system increases the likelihood that a tragedy can be averted before it happens.
Even if it means having a few controllers doing less than they “could” during lower-volume traffic times, this is worth doing. Wouldn’t you rather have us be a little overstaffed, have a few extra procedures that serve as tripwires to potential disaster, and aim to be a little safer? It certainly seems better than hoping a controller working too many things at once will keep up.
When demands are low enough that one controller can readily handle both (ie NOT like we saw the other day), all operations can be conducted on the SAME frequency – ie the ground frequency is not in use during quiet periods ie overnight.
A brief mention on the ATIS would let everyone know. That’s how it’s done in other parts of the world, anyway, and heightens situational awareness for all concerned.
What Higher-faster-unaffordable said! When tower and ground control are combined without combining the frequencies, pilots on one frequency can’t hear what is being said by pilots on the other frequency. That blocks some situational awareness from both pilots, increasing risk without benefit. Combining frequencies is an easy win for safety. Obtaining funding to fully staff control towers seems to be politically unattainable, sadly.
The idea doesn’t address the problem, which is to have sufficient ground/local/flightdata personnel in the tower to make it work. Having every pilot change freqs and contact tower, and then on the other side of a crossing runway, recontact ground to continue taxi, is unworkable. Local/Ground need to coordinate properly and that’s the fix that was missing in this case (due to insufficient staffing, probably, and with the added ASDE issue with the trucks adding more risk to LGA). And eliminating as many crossings as can be done any with is another “fix” (e.g., more perimeter roads for ground vehicles where possible). Tens of thousands of more frequency changes each day on the airport surface, not so much.
I really think the staffing issue is at play here. While the controllers were doing their best, they sure could have another set of eyes on this one.
Having said that, I keep wondering why the guys in the truck didn’t look right before crossing the runway. Every runway I approach, before entering, I always look to see if there is an aircraft on final. Had they done that, they most certainly would have seen the lights of that RJ.
Exactly
Even with separate GND and TWR controllers, GND can be overwhelmed by clearance matters. I was at APF at noon on a Friday and I suppose a whole lotta high-rollers wanted to escape the heat over the weekend because GND was com-jammed with requests. He stayed good-natured but every other phrase was, “no one talk until I talk to you.” Then the conga line of GA jets baking on the taxiway because I guess releases even on a CAVU day were well-spaced. We could have launched five VFR bug smashers for every jet that took off.
Improving voice-less communication like CPDLC ought help the clearance mess, and probably will trickle down to the piston fleet eventually for automated surface clearances so controllers can focus on monitoring more closely rather than talking.
This sounds great…..but, as a point, many (most) contract towers aren’t authorized enough staffing at full staff level to meet that requirement. In fact, some of the busiest contract towers aren’t authorized enough trying to function from 6AM till midnight on six total staff. This is NOT a function of not hiring enough controllers, it’s contract policy.
Now, let’s suppose a facility has full staffing, which permits two controllers at all times…..and I support this strongly. Even this staffing isn’t going to get the job done by Nathan’s proposal. Why? Because that would mean shutting down ops on either ground or flight when a biology break is required. At a REALLY busy airport, there will need to be at least three controllers active at any given time, then. But, what about the “less busy” airports and those less busy times at larger airports? This wouldn’t demand a significant increase in controller staffing, it would probably require doubling the current number of controllers.
Frankly, there are times and places that could easily be handled for short periods by a single controller. But….
Now, let’s talk about the absolutely crazy shift schedule that controllers are forced to work. And what kinds of incentives we could offer to encourage new hiring…. This is an extremely tough and stressful job, and not just anyone can do it. And, frankly, with the attitude toward government employees by the current administration, who would WANT a job there?
At one time, ground control was authorized to use “look and go” when clearing an aircraft across an active runway. It actually worked good, except when it didn’t. It almost didn’t with me at LIT once long ago. I cleared one across at exactly the same time local cleared one for takeoff. Ugly. But, anyhow some years later, “look and go” was cancelled. At that point, the ground controller was then required to ask local if he/she could cross a runway, using specific phraseology. “Cross the fire truck at 4/22 on Delta”. Local was required to approve with exact phrase. I assume that is still required, maybe. However, on a mid shift or when traffic was light, the procedure would frequently get compromised and back to unofficial “look and go”. What happened the other night, don’t know.
One thing I have learned from the accident is the existence of automated Runway Status Lights (RSL) and Take-off Hold Lights (THL) that were installed and operational at the time of the accident. The FAA rules are you cannot cross the runway (RSL) or take-off (THL) if the lights are red, even if you have been cleared by the ground or tower controller. As long as we have people in the loop, bad things are going to happen occasionally no matter how good the technology is or how many people you have in the tower.
In the video, it looks like all the other emergency vehicles had stopped well back behind what looks to be the RSL.
… which would be consistent with the first ARFF rolling across the hold line at exactly the same time as the controller started yelling “STOP!” At that point, he may have thought that, rather than trying to back up and colliding with the ARFF behind him, his best bet was to hi-tail it down the side of the runway until he found a place to get over. He almost made it.
It’s a matter of luck and timing…
Seems to be that everyone is missing the point concerning this tragic accident. The fire truck driver is the one who caused the accident. I cannot fathom why a driver in charge of such a vehicle would venture onto a runway without checking to see if it was clear regardless of who gave him clearance. If he was taking a driving exam he would have failed.
Unless he had an ill-timed comms failure? Or, that because he was on the ground freq he didn’t know that a jet had been cleared to land on the runway he had been cleared to cross? A bee flew up his nose and he ripped his helmet off at a most inopportune time?
I’m not sure what undergirds your confidence that the driver was solely at fault in such a complex scenario. The fact that you “can’t fathom” any alternate cause says little for your imagination or experience, I’m afraid.
Where does the pilot responsibility come into play! Don’t they have windows to look out for possible conflicts regardless of a clearance given?
I cannot understand how you can fault the pilot in this instance. They were cleared to land and were already on the ground rolling out before the truck entered the runway. Thrust reversers deployed and brakes on.
There are so many contributing factors that I’m willing to wait for the NTSB to ascribe “probable cause”, but I doubt that it will be singular. Short of a meteor strike, there’s always more than one.
For starters, the ARFF was crossing the runway at an oblique angle turned away from the landing threshold, so the driver’s right-seater would have had to do the scan for landing traffic.
Did he look? Assuming he did, did he see the landing jet? Expectation bias (they were already cleared to cross) could have made that scan cursory, or the lights of the landing jet could have been missed in the visual clutter.
What did he communicate to the driver? Those suckers are loud; was there a comms issue inside the truck that prevented or delayed the driver from hearing any warning?
The controller saw the deal developing and immediately began repeatedly calling for the ARFF to stop. One of the NTSB’s questions will surely be, “Did the ARFF driver hear any of the “STOP!” calls?” It’s unlikely the driver would have continued across an active runway had they been received, so why did that not happen? Comms issues inside the ARFF. Frequency walked on?
“Success has many fathers, but failure is an orphan.” The NTSB will see to it that this one is not.
The STOP STOP STOP TRUCK 1 STOP call came less than 10 seconds before impact, at which time even if the ARFF driver heard the call and understood, the time required to react would probably not have been adequate. There is definitely debate as to whether the ARFF even thought the STOP calls were for them.
Further supporting the argument that there were multiple factors in this multi-player accident, and anyone who asserts a “lone-gunman” was at fault, simply does not understand the complexity of the problem.
The sad reality is that the two guys who died in the cockpit were the least culpable.
One thing I’d like to find out from the investigation is what frequency was used to give Truck 1 clearance to cross runway 4? If it was on the tower frequency, was it also heard in cockpit of the Jazz flight? That might have given the crew more SA as to what was developing on the ground.
Kirk – I listened to the audio of the accident via RealATC.net and the comms (both ways) between ATC and the ARFF *were* on the tower freq which was also the freq on which the RJ received their landing clearance (which was a “cleared to land #2” clearance over a minute before they actually touched down). Clearly, the tower controller lost “the flick” regarding where the RJ was at the time he cleared the ARFF across the active, although the RealATC audio contains a slight delay between the request from the ARFF and the go-ahead from ATC during which the controller was presumably assessing whether the RJ was past the crossing point. Clearly, he wasn’t, but that fact was missed in the tower cab. My guess is that the RJ crew heard the clearance issued to the ARFF, but it came at a point in the landing sequence at which they had few options. I am sure that there was also a “this cannot REALLY be happening to us!!” denial of reality in the cockpit that made the eventual outcome even more likely.
How about we do the procedure that is required – and sometimes skipped by some who are less than vigilant – in the airline world? One NEVER CROSSES A RUNWAY, and one NEVER TAKES A RUNWAY unless BOTH pilots look outside and after clearing their side, state the words “Clear left.” “Clear right.” ONLY THEN does pilot flying cross the runway or take the active.
Obviously this was the firetruck’s lack of action at LGA.
Like we taught our daughter and now teach our grandchildren: Here’s a street. Whaddya do? Look left, look right, look left, no cars.
If only somebody had looked out the stinkin’ window before taking a runway.
I’m sure there will be a lot of discussion about ATC staffing, but you never hear about how tricky predicting necessary staffing can be. If that shift was reviewed for the three or four previous nights, you might find it ran smooth with no issues at all with the same number of controllers. On the accident night, they had United abort t.o. TWICE! Then try to get back to a gate and while waiting threaten a potential evacuation. That turned what could have been just a busy night into the nightmare it became. Much like hospital ER’s, it doesn’t take much to swamp a well thought out staffing plan.
Plus, the RJ was several hours late in its arrival at LGA; had it been arriving at ANY OTHER TIME, the accident would probably not have happened. One more hole lining up in the swiss cheese model.
Add 200% more checkins for each runway crossing?
We already have pilots with eyes; controllers with eyes; ground drivers with eyes; radios for all the above; ground surveillance radar in the tower; and a set of red lights which warn of a runway in use, and which are not cont
…rolled by any of the above. This was a terrible accident, but it seems like the proposed solution is not really going to help?