Say Again? #62: Too Close for Comfort

  • E-Mail this Article
  • View Printable Article
  • Text size:

    • A
    • A
    • A

We can learn a lot from accidents, but wouldn't it be great to learn from near accidents -- those that don't actually have to be reported? AVweb's Don Brown recently saw some dangerous situations that point out the problems with GPS, and he recounts them in this month's Say Again column.

Say Again?

Whenever there is an accident in aviation, those of us in the business start thinking of all the things that might have gone wrong. We try not to speculate (at least not in public) as we wait for the National Transportation Safety Board (NTSB) to conduct their investigation and let us know what happened.

Most of the time the NTSB's findings are depressingly familiar. Sometimes they have a new twist on an old problem and occasionally they find something unique. Pilots and controllers read the results and discuss them, supporting an informal learning system in the aviation industry. It's easy to get our attention, especially if we feel some connection to the particular incident. It might involve the type airplane you fly or an instrument approach that you've flown before. For controllers it might be the airspace we work or an aircraft that is based at our facility.

For instance, the NTSB recently released their report about the King Air 200 crash at Stuart, Va., on October 24, 2004. It piqued my interest because it happened at a facility I work with (GSO Approach at Greensboro, N.C.,) and it involved an aircraft we worked on a regular basis at Atlanta Center. The "probable cause" was all too familiar -- "controlled flight into terrain" -- but the circumstances leading up to it were interesting. For me, it fit into my belief that non-radar operations, full-approach procedures and an over-reliance on GPS spell trouble in today's world. I urge you to read the report and see what it says to you.

For every accident the NTSB investigates there are hundreds of other incidents that came close to being an accident. While "close" is a subjective term, there is much we could all learn for the circumstances that led up to these incidents. If only we had a way to learn about them.

Too Close

While you mull that over I'm going to tell you about a few incidents I've seen. There probably won't ever be an official report filed about any of them. To be honest, I don't know if some of the pilots involved even realized how close they came to having an accident.

For reasons that escape me, the FAA recently turned the ILS approach around at the Statesville, N.C., airport (SVH). The ILS used to be to Runway 10 but now it's to Runway 28. This has presented all sorts of problems for controllers, but I'll limit myself to the ones germane to this particular case.

First, the final approach course is no longer depicted on our radar scopes. That means that we can't vector for the ILS. In addition, the TEMKY NDB was decommissioned instead of being moved to support the new ILS. I could say that TEMKY was replaced by the PEGTE intersection/waypoint but I believe it's that kind of thinking that led up to this incident. PEGTE, like so many intersections/waypoints these days, isn't displayed on the radar scope. In that most aircraft are equipped with GPS, the majority of pilots proceed direct to PEGTE for the ILS approach. To sum it up, controllers can't vector for the ILS Rwy 28 SVH, so it's a full-procedure instrument approach where most pilots proceed direct to a fix that controllers don't know the location of with any precision.

Strike One

On this particular day, the traffic was pretty heavy into SVH so I took the time to draw PEGTE on the scope. This is a "shade tree" technique that doesn't change any of the rules but it helps me with the timing of multiple approaches. It's easier to provide the correct spacing between aircraft if you know the distance to the initial approach fix (IAF). While the approach is relatively new, I've used it dozens of times before. As a matter of fact, I'd probably used it a dozen times already on this particular day.

"November one two three four five, five miles from PEGTE, maintain three thousand four hundred until PEGTE, cleared ILS Runway two eight approach Statesville."

Duller than dirt, right? The minimum IFR altitude (MIA) is 3,400 and once the pilot was at PEGTE he was on a published segment of the approach. My job as a controller was done. The only difference between this approach and the dozens of others is that I had a close approximation of PEGTE displayed on my scope. The pilot hit PEGTE and turned outbound for the procedure turn, and I decided to look at the approach plate again because I had nothing else to do and I didn't have it memorized yet.

A Closer Look

KSVH (Statesville) ILS-LOC/DME Runway 28 Profile View (click for larger version - 220 Kb)

I looked at the profile view just because I'm curious how the altitudes look without an NDB on the approach. Well, would you look at that? It isn't 3,400 until PEGTE outbound like it'd be with an NDB as the IAF ... it's 3,400 until PEGTE inbound. I looked up at the scope and noticed that I was not the only one who had made that assumption.

The aircraft on the approach was at 2,800 about three miles outside PEGTE. As I was trying to take it all in, the altitude readout clicked down another hundred feet to 2,700.

"November one two three four five, check altitude, it's three thousand four hundred until PEGTE inbound and I'm showing you two miles outside PEGTE."

So much for official ATC phraseology. Knowing what you're supposed to say and being able to say it when you're still recovering from a surprise are two different things.

Of course, by the time we got it sorted out the airplane was over PEGTE and the point was moot. It's real easy to just sluff this off and say, "No harm, no foul." The pilot realized his mistake and, like most of us, he'll be more careful in the future. But the system would be much better off if it wasn't just he and I that were learning something from the mistake.

Strike Two

KHKY (Hickory) ILS Runway 24 Plan View (click for larger version - 120 Kb)

The next incident happened when the Maiden Radar was out again. There was an aircraft inbound to Hickory, N.C., (HKY) from the east. The NOTAM about the radar being out had been issued and GSO Approach assigned the aircraft a heading to join V20 to BZM and then direct HKY. When the aircraft came on my frequency, I advised him to expect the SANFI transition for the HKY ILS Runway 24 approach.

From my perspective as a controller, nothing could be simpler. SANFI is an intersection on V20 where the HKY localizer intersects with V20. The pilot doesn't have to fly a procedure turn even when the radar is out. Just fly V20 to SANFI and go straight in on the localizer. I've been doing it this same way for over 20 years. Evidently, pilots have a different perspective these days. That perspective is GPS.

Just after I told the pilot to expect the SANFI transition he requested direct SANFI. I told him I needed him on V20 because the radar was out. As I've said before, random routes require radar. While the aircraft was still in radar contact, I issued him the approach clearance:

"November one two three four five, ten miles from SANFI, maintain six thousand until SANFI, cleared ILS Runway 24 approach, Hickory.

Just as expected, the aircraft disappeared from the radar a few moments later.

"November one two three four five, radar contact lost, report passing SANFI."

How Close?

The aircraft was only a few miles from SANFI so it shouldn't have been but a minute or two before it was by SANFI. After SANFI, I planned to have him report by MIRTY intersection, get an "altitude leaving" report from him and then I'd switch him to the Tower. This is the process controllers use when it's non-radar. You keep up with the times a pilot reports passing various fixes -- and the altitude reports -- and you update your mental image of where all the airplanes are.

After talking to a few other airplanes I realized that I still hadn't received a report from the pilot passing SANFI. Maybe the frequency was busy and he couldn't get a word in.

"November one two three four five, say position."

"November one two three four five is three miles from Barrett's Mountain."

That doesn't quite make sense. SANFI is 13 miles from BZM. After SANFI he'd be on the localizer and it'd be unlikely he'd give me a DME off of BZM as a position report. So, I ask another question:

"November one two three four five, say altitude."

"November one two three four five is level six thousand."

Now I know we have some major confusion. If he's three miles from BZM and on the localizer, he ought to be at (or descending to) 3,400. To make a long story short, I believe the pilot was thinking he had to fly V20 all the way to BZM and then turn around and go back to SANFI to shoot the approach. Without a report from the pilot, we'll never really know.

There was very little chance that this particular incident could have resulted in an accident. Yet, I still can't help but be concerned that it resulted in a navigational error this large in an approach environment. Working in an en route environment, I'm used to seeing a certain amount of inattentiveness by pilots en route. Controllers recognize that the en route phase of flying is mostly boring for pilots. But these errors are occurring during the approach phase where a pilot is (supposedly) operating at an increased level of awareness.

Strike Three

Let's look at one more incident. I was working another aircraft inbound to an uncontrolled airport in the mountains, Banner Elk, N.C., (NC06). The MIA in the vicinity of the airport is 7,500 and I assigned the aircraft 7,500. Very near this airport, the MIA increases to 8,000. In addition, it is near another facility's border. Like almost all aircraft these days, this aircraft was proceeding direct to the airport using GPS. But because of the higher MIA and the border, I didn't want this aircraft wandering around, so I assigned the aircraft a heading:

"November one two three four five, fly present heading, vector to Banner Elk. Report Banner Elk in sight.

The pilot acknowledged the heading assignment. Three minutes later I looked up and the aircraft was circling the airport.

If an approach controller at Atlanta Approach had told this pilot to "fly present heading, radar vectors Atlanta," would this pilot have started circling Atlanta-Hartsfield -- looking for a hole in the cloud deck -- without a clearance to do so?

While I'm here, I'll restate another observation I made in previous columns that I believe is the flip side of this same coin: If a pilot has been cleared direct to an airport, why do I see them turning themselves direct to a fix other than the airport? They don't have a clearance to fly direct to the NDB, IAF or any other waypoint/fix.

GPS Strikes Again

The common link in all these incidents is GPS/Advanced NAV. (For simplicity's sake, I'll just stick with the term GPS to refer to Advanced NAV for the rest of this article.) Before my detractors get too far along in their train of thought, let me state the obvious: GPS is a huge improvement over VORs and NDBs. The accuracy and the lower cost have revolutionized air navigation.

Most people, if asked, would say that GPS greatly increases their situational awareness. "Situational awareness" is a somewhat fuzzy term to start with but it's a very useful term so I'll use it. I want to advance the argument that GPS hasn't improved situational awareness. It could, but it hasn't. It has increased "positional awareness." I may be splitting fuzzy hairs but I think they need to be split. Perhaps a better way to say it would be that GPS has greatly increased positional awareness while unintentionally decreasing situational awareness.

I'd like for you to think about this subtle difference in the three situations I described above. For the approach at SVH, the pilot knew where PEGTE was with great accuracy and ease. That positional awareness seems to have come at the expense of his situational awareness, though. He wound up 800 feet too low, too soon.

For the pilot shooting the ILS approach to HKY, while I feel the pilot knew exactly where SANFI was, he wound up doubling back to the fix in a non-radar environment. Flying a transition from an airway to an approach isn't new. It's old. Really old. Yet somehow the use of GPS has resulted in a different mindset that led to confusion.

For the aircraft inbound to Banner Elk, the positional awareness afforded the pilot by GPS seems to have overcome his situational awareness -- the fact that he had been assigned a heading to fly by ATC. The GPS saying the airport is "here" seems to have overcome the pilot's memory that he was told to fly "there."

None of these examples resulted in an accident. It doesn't take a lot of imagination to understand that -- given slightly different circumstances -- the underlying causes could have resulted in an accident. Sometimes terrain changes rapidly in very short distances. There are always other aircraft about. If you unintentionally cross over into another controller's airspace, there's no telling who else might be out there. Making an unexpected course reversal in an unexpected place -- when it's non-radar -- is almost as spooky as being 800 feet too low on a precision approach.

A is for Aeronautics

The question remains, though. How can we -- all of us -- learn from the incidents that don't result in accidents? How do we learn before it's too late? The best answer I know of is (still) the NASA Aviation Safety Reporting System.

I urge you to go to that site and learn something. Read the back issues of Callback. Go to the database page and read some of the database "sets" or go to the database itself and search for something to learn. Search for the type of aircraft you fly or search the database to see the incidents at your local airport.

And above all, go to the reporting forms page and print yourself a form. Put it in your flight bag so it will be handy the next time you make a mistake others could learn from. Fill it out and mail it in.


Before I go, I must tell you about an experience I had after I'd finished writing this article. It was my first experience with GPS. No, it wasn't in an airplane. It was in my wife's new car. We got stuck in a typical Atlanta traffic jam and I thought I'd test out the GPS system.

I got off the Interstate and we used the GPS to navigate some unfamiliar back roads. It was amazing. I could see how we were paralleling the interstate on the moving map. And the voice commands, telling us when to turn, were a tremendous help when we came to the town square we had to go around. Right up until I turned onto a one-way street -- going the wrong way.

Have a safe flight.

Don Brown
Safety Representative
National Air Traffic Controllers Association

Want to read more from air traffic controller Don Brown? Check out the rest of his columns.