NATIONAL TRANSPORTATION SAFETY BOARD

THE INVESTIGATION OF KOREAN AIR FLIGHT 801,
B-747-300, AGANA, GUAM
AUGUST 6, 1997

Ballroom A and B
Hawaii Convention Center
1833 Kalakaua Avenue
Honolulu, Hawaii 96815
Thursday, March 26th, 1998
8:00 a.m.

NTSB Board of Inquiry Members

Technical Panel Members

Public Information Officer

General Counsel

Parties to the Hearing

A G E N D A

AGENDA ITEM

Testimony of Nelson Spohnheimer
National Resource Engineer for Navigation
Federal Aviation Administration
Renton, Washington

Testimony of Captain Paul Woodburn
British Airways
Chairman, ICAO, CFIT Steering Committee
London, England

Testimony of Donald Bateman
Chief Engineer, Flight Safety Systems
Allied Signal, Inc.
Redmond, Washington

Testimony of William Henderson
Manager
Western Flight Procedures
Development Branch
FAA Western Pacific Regional Office
Los Angeles, California

Testimony of James Terpstra
Senior Corporate Vice President
Flight Information Technology and
External Affair
Jeppesen Sanderson, Inc.
Englewood, Colorado

Testimony of Captain Wallace Roberts
Former Chairman, ALPA CHIPS Committee
Air Line Pilots Association (ALPA)
Herndon, Virginia

Conclusion

P R O C E E D I N G S

8:00 a.m.

CHAIRMAN FRANCIS: Mr. Feith's watch indicates that it's now 8:00. So, I think we'll get started, and hopefully with an eye on making certain that we do our usual comprehensive job. If we can move along, maybe we won't still be here at late in the day this afternoon.

Our first witness is Nelson Spohnheimer, National Resource Engineer for Navigation at the FAA in Renton, Washington.

Whereupon,

having been first duly sworn, was called as a witness herein and was examined and testified as follows:

TESTIMONY OF NELSON SPOHNHEIMER
NATIONAL RESOURCE ENGINEER FOR NAVIGATION
FEDERAL AVIATION ADMINISTRATION
RENTON, WASHINGTON

MR. SCHLEEDE: Please give us your full name and business address for the record.

THE WITNESS: Yes. Good morning. My name is Levi Nelson Spohnheimer. I work for the FAA at the Northwest Mountain Region Headquarters in Seattle, 1601 Lind Avenue, SW, Renton, Washington 98055.

MR. SCHLEEDE: Thank you. And what is your position at the FAA?

THE WITNESS: Well, my title is National Resource Engineer for Navigation, which -- which means that I work on a wide variety of technical topics related to all kinds of ground-based navigational aids and their flight testing.

MR. SCHLEEDE: Would you give us a brief summary of your education, training and experience that qualifies you for this position?

THE WITNESS: Surely. I have an electrical engineering degree from Iowa State University. I worked for about six years in industry for Texas Instruments and Motorola as a radio frequency design engineer. During that time, I became system engineer on an instrument landing systems contract, and as a result, I joined the FAA. I've been working on ground-based nav aids of all types for about 24 years.

MR. SCHLEEDE: Thank you very much.

Mr. Phillips will proceed.

MR. PHILLIPS: Good morning, Mr. Spohnheimer.

THE WITNESS: Good morning.

MR. PHILLIPS: Have you had any accident investigation experience in your career?

THE WITNESS: Well, yes, I have. I'm -- I'm the Northwest Mountain Region accident representative for airway facilities, and I work on various national accidents, typically those having navigation issues.

I've worked on the litigation of a number of cases, and I've served on the Air Force Board for the Bosnia accident.

MR. PHILLIPS: Okay. Most of your experience then has dealt with the ground-based side of the equipment?

THE WITNESS: In general, that's correct. I -- I spent a lot of time with the airborne flight testing organization, but most of my work is on the ground equipment.

MR. PHILLIPS: Could you describe a typical work day for yourself?

THE WITNESS: Well, fortunately, it varies quite a lot. I travel extensively, about 40 weeks a year. So, each week is different. But in a given month or two period, I might teach a technical class or seminar, do some trouble-shooting work on signal and space problems with ground-based nav aids, visit two or three companies who have applied for FAA approval for their nav aids equipment, write some technical papers.

I serve on a couple international civil aviation organization committees that deal with standards and testing of ground-based nav aids.

MR. PHILLIPS: Okay. Have you been present the last two days during the testimony in the hearing, and are you familiar with the issues in this hearing?

THE WITNESS: Yes, I have, and I am.

MR. PHILLIPS: Okay. And specifically, I realize that your expertise covers a lot of areas, I'd like to address my questioning today in the areas of the instrument landing systems, and along those lines, I'd like to ask you just a few questions about what is an ILS. Let's lay a little foundation for what is an instrument landing system, how does it work. Go ahead.

THE WITNESS: Okay. An instrument landing system is a ground-based electronics system composed of about six subsystems that provide lateral and vertical guidance and fixes or rough knowledge of position to the pilot along the approach path to an airport.

MR. PHILLIPS: Would you -- would we like to go ahead and put up Page 6 of Exhibit 9-E, Teddy? Would this help in your discussion?

THE WITNESS: Well, yes, thank you. This is the simplified but sufficient diagram of the nature of the needle indications that are provided to a pilot while flying an instrument landing system approach.

The needle, as you can see in the bottom right-hand corner of the -- of the picture or the indicator, rather, consists of two needles, a fly right/fly left and a fly up/fly down, and the antenna system on the ground is arranged in such a way that these needles deflect proportionately more and more as the aircraft departs more and more from the desired course or glide path.

The system operates by transmitting two tones, much like two notes on the piano, and these tones are arranged to be equal in signal strength on the desired path, and -- and as the airplane moves from the desired path, the two tones become unequal in magnitude, and -- and it is that inequality that moves the needles on the cockpit indicator.

MR. PHILLIPS: Is this -- is this the standard ILS system used around the world? Are there any differences in the design?

THE WITNESS: No. This -- this basic character is -- has been standardized worldwide for nearly 50 years.

MR. PHILLIPS: Okay. Speaking of standards, are there technical standards that dictate the design requirements for ILS components?

THE WITNESS: Yes, there are a number. Internationally, the signal and space is defined by the International Civil Aviation Organization in a document called "Annex 10". The standards -- the standards are listed in the manner that define very fully the signal and space characteristics.

Receivers, which must use that signal and space, have their characteristics defined by, in general, two organizations, RTCA in the U.S., which is the Radio Technical Commission for Aeronautics, and a European equivalent called Eurocae, E-U-R-O-C-A-E. These bodies are -- are consortiums of manufacturers in general and regulatory agencies, and their standards define how the receiver will react to the signal and space that's defined by ICAO.

MR. PHILLIPS: Okay. Do the FAA requirements require these standards to be met before they're installed on airplane or ground-based equipment?

THE WITNESS: Yes. For -- for most operations, certainly air carrier operations, the -- the receivers must meet what's called a technical standard order, a TSO, which FAA publishes. It provides a regulatory trail to the RTCA standards in most cases. So that an approved installation on an airplane of an instrument landing system receiver must meet the applicable RTCA document.

MR. PHILLIPS: How long have instrument landing systems been in use?

THE WITNESS: Difficult to say precisely, but the early development occurred roughly at the beginning of World War II, and the system that we know today was pretty well standardized by the end of World War II.

MR. PHILLIPS: Okay. Have there been enhancements or improvements over the years into that system?

THE WITNESS: Well, yes, on the ground side. Although the basic signal generation system has been pretty much standard, the antenna systems that attempt to keep the signal of high quality, straight, with no variations along the approach path, have had to get more and more advanced due to the encroachment of hangars and other reflecting sources on or near airports.

So, in -- in the main, the science of instrument landing systems is the science of antenna systems on the ground side.

On the airborne side, of course, as we went from tubes to transistors to integrated circuits and now software-based or receivers that contain software, there's a continual advancement in the performance, but the general description and the general way in which all ground and airborne systems behave has remained unchanged.

MR. PHILLIPS: In regards to some of the RTCA standards that control the design or specify the design for this equipment, specifically DO-131, 132, 192 and 195, can you elaborate on -- on your opinion or assessment of the differences in these standards over the years?

THE WITNESS: Surely. Two of those were published in 1978 and defined how localizer and glide scope receivers should behave, and the other two were published in the mid-'80s, I think 1986, and were updates to address the changing environment in which aircraft operate.

For example, the occasion of other transmitters, paging systems, cellular radio systems, tv systems and so on has meant that receivers have to be able to operate in more and more demanding environments. As these installations encroach around airports, the frequency congestion gets higher.

So, one of the areas about receiver design that has received a lot of attention in the last four or five years is increased immunity to such out-of-band signals, and, of course, the software is a new -- relatively new change in airborne equipment, and one of the more recent updates deals with software quality assurance.

MR. PHILLIPS: Okay. Are you aware of any accidents or incidents where ILS system components, ground-based side, because of your experience, have been an issue?

THE WITNESS: No, I'm not. I have worked on a number of lawsuit cases as a witness, and to my knowledge, no instrument landing system has been found causative for an accident.

MR. PHILLIPS: You heard the testimony, I believe, in the beginning day of the hearing, the crew of Korean Air Flight 801 commented several times about a glide scope signal or at least the glide scope flag, glide scope operation, when the -- we know -- we know that the glide scope equipment wasn't present at the time, the transmitter.

Would you like to comment on that in general terms?

THE WITNESS: Well, yes. As you say, the -- the intended glide scope signal had been removed for service to replace its shelter and was out of service for about a month prior to and after the accident.

The pilot would normally be warned that a signal is not present by the presence of a flag, a warning flag, that indicates that something about the receiver system or something about the ground system is abnormal, and one has to assume that these remarks had to do with the presence or absence of flags.

There are enough remarks in the record that I have to conclude that there must have been some sort of flag activity coming into view, disappearing from view, some time during the approach.

MR. PHILLIPS: Is that unusual in lieu of the fact that we know no transmitter was present?

THE WITNESS: Well, no. When we have an empty channel, many of these potential external sources of noise and unintended signals, which are normally too weak to be heard, can be heard, and it's fairly common when we test airborne flight tests, instrument landing systems, and we turn off the localizer or the glide scope that -- that we record on our instrumentation intermittent indications of flag and needle activity, and as a result, the aviation community relies on notices to airmen as a procedural means to advise everyone that the channel is empty.

MR. PHILLIPS: Would you expect these -- this flag movement to cover a time period that would indicate to a crew that the signal may be valid?

THE WITNESS: Well, no. The typical case of finding some sort of activity on instrumentation is very short duration, intermittent, and -- and pilots usually refer to these brief movements of the flag as flag pops.

For a crew or a pilot to conclude that a signal is on the air and flyable would probably require the flag to remain in a static condition for 10 or more seconds perhaps.

MR. PHILLIPS: Is there any indication in your mind in the transcript, the CVR transcript, to indicate that the length of time these flags may or may not have been in view?

THE WITNESS: Well, the -- the individual comments, of course, do not convey much information about the duration of any flag activity, but I would conclude that there must have been enough absence of flag for the crew to occasionally decide that the system was on the air when in fact it wasn't.

MR. PHILLIPS: If the flag moved out of view, would you have expected to see a needle deflection of any sort, a fly-up or fly-down positional command?

THE WITNESS: Well, on an empty channel, that's very statistically hard to determine. The nature of the various interference or noises, electrical noises, that might cause the flag to move is pretty random, and, so, some of those will cause a quick deflection of the needle, returning it to zero. Others might deflect a needle for a short time. It is quite random in the general case.

We have many recordings from our flight test organization that shows what most people would call erratic needle movement.

MR. PHILLIPS: Can you elaborate a little on the flight test of an ILS system? What's done, and the frequency, and --

THE WITNESS: Yes. In the U.S., instrument landing systems are flight-inspected on a periodic basis. The -- the period ranges from a few months to about 10 months at the maximum.

During each of these flight tests, the alignment of the localizer and the glide path, the amount of needle deflections when the aircraft is off the path, and the actions of the ground-based monitoring system that removes the signal from service when it exceeds certain standards, are all tested and recordings are made. Every other flight inspection is a brief one, might take 30 to 45 minutes. The alternative flight inspections typically take several hours.

MR. PHILLIPS: These flight inspections are conducted with specially-instrumented aircraft or ground-based or --

THE WITNESS: That's correct. I'm speaking about the airborne testing. The aircraft are equipped with quite a lot of unusual avionics and recording capability that provide engineering quality measurements of the signal characteristics.

MR. PHILLIPS: Okay. Could you address flight testing at the Guam Airport; specifically, the post-accident testing that may have been conducted on the system?

THE WITNESS: Yes. We -- we, of course, have a policy that after accidents, any ground-based navigational aids that may have been involved are -- are flight tested as quickly as feasible after the accident, and, so, of course, there was a flight test of those components of the ILS that were in service at the time of the accident, and everything was found normal.

MR. PHILLIPS: Okay. Back to an earlier discussion of the ILS system, we didn't talk about the marker beacons. Would you give us a general description of what a marker beacon is, and what it is to an instrument approach?

THE WITNESS: Surely. A marker beacon is a small and fairly simple ground-based transmitter system that transmits an upward-directed antenna pattern through which the airplane flies on the approach. It causes a separate receiver in the aircraft to light a particular light, different-colored lights, for the different markers that are usually installed on an approach.

The outer marker, inner marker, and middle marker would be the full complement for a high-precision landing system. Each one has a separate light on the instrument panel. So, for about five to 15 seconds, as the aircraft flies through the antenna pattern of each marker station, the associated light will illuminate.

MR. PHILLIPS: Testing of the marker beacons is a part of the flight check?

THE WITNESS: That's correct. The -- the lineal distance along the flight path, the time for which the light is illuminated, is tested and set to a specific value.

MR. PHILLIPS: Is there anything to alert the ground control tower or ATC specialist that a marker beacon system is inoperative?

THE WITNESS: That varies with the installation. In the general case, in the U.S., we do remotely monitor the status, on-air/off-air status of -- of all the components of an instrument landing system.

Certainly for Category 2 and Category 3 higher-precision systems, that is a requirement. For Category 1 systems, such as at Guam, it's not uncommon for the outer marker and sometimes the middle marker to not have remote monitoring because the absence of -- because of the absence of communications lines, phone lines, being available to remote this indication to air traffic control.

So, at Guam, the outer marker is not monitored. The remote status is not monitored.

MR. PHILLIPS: Would -- would you consider an inoperative marker beacon -- would the ILS be operational with an inoperative marker beacon?

THE WITNESS: Yes, in most all cases. It depends upon the design of the instrument approach procedure, but in the general case, the outer marker absence can be substituted with DME or radar vectoring or a compass locator.

So, it's fairly uncommon that the absence of the outer marker eliminates an instrument approach, an ILS approach.

MR. PHILLIPS: Okay. Backtracking just a little bit on your comments about flag pops, do you -- in your view, in your opinion, do you believe that there's -- there's appropriate FAA guidance regarding flag movement on empty channels, I guess specifically in regards to the airman's information manual and flight training practices?

THE WITNESS: Well, I think so. The -- the airman information manual, of course, describes the situation of navigational aids that are off the air. For example, in the U.S., we have perhaps in round numbers 100 instrument landing system approaches which are based on a localizer-only installation. No glide scope has ever been installed.

So, it is common that pilots have to deal with either a glide scope that's been installed being temporarily out of service, or a glide scope that was never installed presenting an empty channel to every -- every airplane on approach, and therefore the aviation community again, as I said earlier, relies on procedural methods, such as notices to airmen and ATIS announcements, to advise pilots that -- that a particular navigational aid is out of service.

MR. PHILLIPS: If a crew was advised that the glide scope was unusable, do you believe that there's any duration of signal long enough to decide that the approach to the glide scope would be flyable?

THE WITNESS: I'm sorry. Would you perhaps restate that?

MR. PHILLIPS: In your understanding of -- of the instructions to the flight crews in the airmen's information manual, is there any period of time -- if

-- if the approach was -- the glide scope was inoperative or unusable, would there be any duration of flag out of view that would be considered enough to consider the -- the source valid?

THE WITNESS: I guess I'll have to assume that you mean if there's a -- if it's announced that the system is --

MR. PHILLIPS: Yes.

THE WITNESS: -- unusable?

MR. PHILLIPS: Yes.

THE WITNESS: Well, if it's announced, and a notice to airmen has been issued, then I think it's quite clear that no period of flag activity, present or absent, warrants use of the navigation signal.

One reason this must be the case is that even though a glide scope or a localizer may be radiating during periods of ground maintenance, we're required to issue a notice to airmen, and during a period that may last for several hours, the system may radiate signals that appear normal, signals that may be flawed.

The various sorts of testing that must be done on a routine basis for ground maintenance result in signals which, from the pilot's point of view, may appear to be valid. A flag would be out of view. A needle would be deflecting in either normal or abnormal methods or manners, however, and therefore the -- the procedural method of advising the pilot not to use the indications is -- is critical.

CHAIRMAN FRANCIS: Greg, could I interject a question here?

MR. PHILLIPS: Sure.

CHAIRMAN FRANCIS: In a situation where you have a glide scope, a fully-operative ILS system, I assume that the glide scope is subjected to remote maintenance monitoring of some sort that you've got the --

THE WITNESS: That's correct. I think you're perhaps referring to what we call integrity monitoring.

CHAIRMAN FRANCIS: I'm dated.

THE WITNESS: There are -- there are three types of monitoring. One is physically present at the transmitter site, and that integrity monitor will turn the transmitter off any time the signals exceed the international standards.

CHAIRMAN FRANCIS: And when that happens, how does the FAA deal with notification of the pilot community?

THE WITNESS: Well, we issue a NOTAM, a notice to airmen, as soon as we're aware that the system is off the air.

CHAIRMAN FRANCIS: And then ATIS and ATC will --

THE WITNESS: That's correct. Depending on the airport, within a short time, -- well, air traffic control will verbally announce to every arriving pilot until such time as the NOTAM or the ATIS recording has been made accurate.

CHAIRMAN FRANCIS: I think that Mr. Phillips' questioning on this, how do you -- how do you make certain that pilots are sensitive to the fact that when they're getting the NOTAM, the controller clearance or whatever it is, that they must ignore any flag activity in the cockpit is -- is one that it certainly would be interesting for the FAA and the international community to pursue, how in training, in the AIM or wherever it is, that -- that we -- we emphasize that enough so that you at least minimize the distraction factor.

THE WITNESS: Certainly. The airmen's information manual and -- and ground school in the general case addresses these issues, although I don't have any oversight knowledge about how -- how thorough that is.

CHAIRMAN FRANCIS: Okay.

MR. PHILLIPS: As part of this flight testing and ground testing of the equipment, are the FAA technicians who perform these tests and review them specially trained or certified?

THE WITNESS: Yes. The ground technicians who maintain an instrument landing system must earn certification credentials by attending a theory class or -- or taking a bypass examination, receiving some on-the-job training, and demonstrating proficiency in a performance examination administered by someone who is already certified, and once the credentials are earned and an assignment to maintain a facility is made, then the national ILS maintenance handbook defines the types of tests, the periods for the tests, the frequency, in general provides the guidance necessary for the technician to periodically test and make a judgment that the system is safe to leave in operation.

MR. PHILLIPS: I'd like to go back to the area of needle movements and flag pops and the potential for those kinds of activities.

Can you describe some of the signals that would potentially cause the flag to move or the needle to deflect, the source of the signal?

THE WITNESS: Okay. Certainly. I mentioned that the -- that the ILS operates by transmitting two tones, and the difference in the signal strength of those tones is what deflects the, in the case of a glide scope, the fly-up and fly-down needle.

So, that means that the receiver has some circuits in it which are looking for those two particular tones, filters that --

MR. PHILLIPS: Would this be a good point to put up Exhibit 9-G?

THE WITNESS: Perhaps, --

MR. PHILLIPS: This was --

THE WITNESS: -- if that's the --

MR. PHILLIPS: Yeah. That's the --

THE WITNESS: -- diagram.

MR. PHILLIPS: -- schematic. For the benefit of the tables, this exhibit was added this morning. It's a one-page aid.

THE WITNESS: Yes. This is a diagram of -- at the most basic level of an ILS receiver on the top half of the -- of the view there.

The filters in the top center labeled 90 and 150 are those filters that are looking for these two particular tones that deflect the needle, and the large circle labeled CDI, course deviation indicator, in the case of a glide scope, for example, is the needle that -- is the meter that the pilots look at, the fly-up and fly-down indication.

So, the fly-up/fly-down needle is an indication of the difference in strength of those two tones, and the difference will be zero, and the needle will be centered when the two tones are equal, and as I mentioned earlier, we -- we go to great lengths to arrange the antenna system on the ground so that those signals are equal at the three-degree glide path.

Now the flag circuit, the other indication that the pilot sees, is driven by a signal which is the sum of the two circuits or the two signals. As long as the 90 and 150 signals are both present at sufficient strength, the flag will remain out of view.

So, the pilot looks at a different signal, which is the fly-up/fly-down, and at a sum signal, although he probably is not aware that it's a sum signal, that activates the flag.

MR. PHILLIPS: Now I -- the localizer works in the same manner as the glide scope, just turned off axis?

THE WITNESS: That's correct. On a different channel, we transmit the same two tones with an antenna system that assures that the tones are equal in signal strength on the runway extended center line, and that drives another needle which has fly-right and fly-left movement, and that needle stays centered again when the two tones are equal in strength, and a separate flag for the localizer is driven by the sum of those two circuits or two signals.

MR. PHILLIPS: So, then when a flight crew dials in a frequency for the instrument approach, they're actually tuning two frequencies?

THE WITNESS: That's correct. The -- the published frequency of the instrument landing system, for example, 110.3, is that of the localizer. The glide scope is paired in a pre-defined way so that the pilot need not also specify this second frequency, but two receivers are being set up on two different channels by that one action of setting 110.3.

MR. PHILLIPS: I see on the bottom of your chart, you have two -- two peaks there that say filter response versus frequency. Would you like to discuss that?

THE WITNESS: Well, yes. Because you asked earlier about what sort of signals could cause the flag in particular to move, we have to know a little bit about the filters that drive that flag circuit.

The bottom figure shows in a general sense how the output of the filters varies for a constant input signal of differing frequencies. To use my two notes on a piano analogy, if you were to play five or six notes on the piano centered around the 90 hertz frequency, only the one that corresponded to 90 would produce, say, a one-volt output of the filter, and as you played other notes at the same level of volume, because they're not at 90 hertz, not at the center of that frequency response for the filter, less and less of the equal -- equal amplitude input signal would be output.

So, as long as the ground station transmits only 90 and only 150 signals, these filters, the 90 and 150 filters that feed the fly-up and fly-down needle and the flag circuits, output equal amplitude signals when the airplane is on course and on path.

If the channel were empty, no ground station transmitting, no intended ground station, and some other signal, for example, a two-way radio with someone speaking on it, should somehow get through the frequency-determining circuits, then those portions of the signal that contain 90 and 150 tones, those portions of the voice, for example, or a music program would still get through those filters and could cause the -- the two needles, the sum and difference needles, to deflect in brief ways.

My voice, for example, contains 90 and 150 hertz components. Music contains frequencies in those ranges. So, depending on the shape of the filters response, which varies from receiver to receiver and from manufacturer to manufacturer, the flag and cross-blender circuits would see varying amounts of intermittent deflections, depending on the content of this spurious signal. As long as it contains 90 and 150 components or frequencies close to them, there's a potential that the needles will deflect.

MR. PHILLIPS: So, then would the -- using that discussion, would the most effective filter be one that had the steepest slope about 90 and 150 hertz points?

THE WITNESS: Yes. When -- when -- when you build the filter for any purpose, you want it to be as selective as possible or as reasonable. The two general curves that I've drawn there are somewhat typical. As -- as technology improves and costs of circuits get lower, it's more common to see narrower and narrower response curves. So that only frequencies very close to 90 and very close to 150 get through to the sum and difference indicators.

MR. PHILLIPS: Then would the effect of this be fewer erratic needle movements and flag movements?

THE WITNESS: That's -- that's correct. In the general sense, the -- the newer the receiver, the sharper the filters, the less often a pilot would see short duration flag pops and needle movements from an empty channel.

MR. PHILLIPS: Assuming we had an empty channel, if we had an intermittent flag, what would the needle be doing or what -- what would you expect it to be doing?

THE WITNESS: Well, for the flag to move, that means that the sum of the output of the two filters has to exceed some threshold that's been previously set.

The flag, of course, cannot tell whether the output from the 90 filter or the 150 filter or both are contributing to the signal that moves the flag. So, it's not possible to say in the general case whether the CDI will stay centered in the case of equal amounts of 90 and 150 or deflect up or down or right or left.

If the external undesired signal was composed of music, for example, the base notes in the music would vary. They wouldn't always be 90 or 150, and therefore if there were enough signal getting through the filters to move the flag, sometimes the needle would deflect up or right, sometimes it would deflect down or left. It's just very difficult to say.

But in the general case, it's random because voice and music and most signals that are transmitted by radio systems do not have 90 and 150 as an intended information source, and therefore those components that happen to be at 90 and 150 are time-variant.

CHAIRMAN FRANCIS: Could I interject a question here? If it's possible, could you characterize the relative sophistication or modern -- how modern the -- the receiver in KAL-801 was in terms of the narrowness of peaks?

THE WITNESS: Yes. I would -- I believe -- I would say that the KAL receiver was fairly typical for recent receivers. There are newer and sharper filtered receivers available, but it is -- the filter response characteristics of that receiver are pretty common. Quite a few other models from various manufacturers have similar characteristics.

The shape of those filters is defined by something called Q, a quality factor, and to get a high-quality factor in a very narrow filter shape takes some more components or some software in the general case. Most of the manufacturers use pretty similar techniques.

As the receiver model generations change over time, the filters typically get narrower, just because it's convenient and cost-effective to make them so, but there are many receivers in service, like the KAL receivers.

MR. PHILLIPS: Okay. One step back here in your description of the deflection without an intended signal, would we need a fairly constant tone then, either a 90 or a 150 hertz range, to cause a steady needle deflection in the absence of a normal glide scope.

THE WITNESS: Yes. Whatever type of signal gets through those filters, it would have to have -- the amount that got through the 90 filter and the amount that got through the 150 filter would have to be fairly constant, so that the difference between the two is constant, and the needle would deflect to a consistent value.

MR. PHILLIPS: In looking at your example of filter response versus frequency on the bottom of the chart, it would appear that approximately halfway in between the 90 and 150 hertz frequencies, at about 120 hertz, the filters would be the least selective, is that true?

THE WITNESS: Yes, that's correct. Where those two responses cross, which can be 120 or 122, it varies a little with the model number, but it's approximately 120, a single tone of that fixed value would get through the filters equally well and would result in, if it were strong enough, a centered needle.

MR. PHILLIPS: Okay. That leads us to a discussion regarding some post-accident testing conducted by Korean Air Lines.

Have you been briefed, and are you aware of those tests and results?

THE WITNESS: Yes, I have.

MR. PHILLIPS: Okay. Would you like to summarize those or would you like me to?

THE WITNESS: I'll take a crack at it.

MR. PHILLIPS: Okay.

THE WITNESS: The Korean Air Lines test basically said what -- what type of signal could cause the flag to disappear from view and cause the CDI to remain basically centered, and -- and since all of us in the business are aware of these filter shapes, as you pointed out, if you had a signal on channel that had in this case 120 hertz modulation, a single tone, it wasn't an ILS signal but it was some other signal, and if that tone were strong enough, you notice that the response of the filters at 120 is rather low, but if the strength of the 120 signal were strong enough, the music were strong enough, the voice were strong enough, for example, then the signal that gets through both filters and is summed in the flag circuit might be sufficient to cause the flag to move.

So, they bench tested such a scenario, a signal generator with modulation of a 120 hertz, quite strong, roughly twice as strong as the typical glide scope 90 and 150 tones, and -- and found that on a variety of receivers, they were able to cause the flag to disappear from view.

Because the filters have a roughly equal response at 120, when the flag disappeared from view, the -- the cross pointer fly-up/fly-down indication was roughly centered, and it would vary from receiver to receiver because the filters are not identical at 120 in every case, but over a wide range of manufacturing choices, most of the receivers have an equal response at approximately 120.

So, -- so, they found out of six different models of receivers from several different manufacturers, four of them, those with the broader filter characteristics, would allow the flag to disappear from view, and two of them with narrower filters left the flag in view.

CHAIRMAN FRANCIS: Any indication of how long that might be -- disappear from view than the -- than the less-precise ones?

THE WITNESS: Well, of course, their tests were static with a continuous signal from a test generator, just to show that the receivers would indeed respond if such a channel -- such a signal were on channel. So, these were -- so far, I've described just bench tests.

CHAIRMAN FRANCIS: I assume we're getting to that.

MR. PHILLIPS: Yeah. Do these results surprise you in any way? Are they what you would expect?

THE WITNESS: They're what I would expect, given the nature of receiver design.

MR. PHILLIPS: Okay. Based on -- on these tests and -- and what you've seen and the testimony this week or what you've heard, do you believe that the warning flags are adequate to protect from interference or -- or spurious movement?

THE WITNESS: Well, no. This -- this type of circuit is intended to warn of failures in the ground ILS station or -- and in the receiver and -- and does not address other types of signals which may have 90 and 150 components.

Obviously any type of signal that's on channel, instead of intended ILS station, if it has the right characteristics in the audio, music and voice and so on, this type of flag circuit, which is used extensively, cannot discern the difference between the intended ILS signal and an extraneous one that has the right characteristics that last long enough.

MR. PHILLIPS: Along those lines, at an instrument landing system location, how do we design or how does the FAA protect the local environment so that those tones and frequencies are predominant?

THE WITNESS: Well, the Federal Communications Commission, which, of course, manages the spectrum in the U.S., has granted to the FAA the management of those bands of spectrum that -- on which the ILS operates.

So, in the general case, of course, we assign instrument landing systems so that any two which are on the same channel are sufficiently far apart that a single aircraft cannot receive two of them at one time.

As far as out-of-band signals go, such as paging transmitters and all sorts of personal communications devices, any time someone is going to construct a station within about four miles of an airport, we have a requirement that they notify us and obtain approval for installation of those stations.

In my region, for example, we see about 30 of these applications a week, and each one is examined for its signal strength, its frequency, its potential to affect radar systems, microwave systems, instrument landing systems, and so on.

So, in that sense, we have a regulatory control over how close and what nature of transmitters are installed close to an airport. So, as long as all of these emitters operate in the way they are intended, the -- the frequency band can be kept clear of non-ILS signals.

MR. PHILLIPS: You noted that in the way they were intended. Does that imply that there's a possibility that an unintended operation could have an effect?

THE WITNESS: Well, surely. Just like any -- anything that we own, like a car or a microwave oven, after some time, transmitters may degrade or fail in ways that cause them to transmit on incorrect frequencies or have incorrect characteristics, and when that occurs, there is a potential in -- in any radio-type system for other systems to be affected.

So, the protection of the navigation frequencies for this condition is basically a reactive one. There's no way to predict when to continue picking on the paging folks, for example. There's no way to predict when a given transmitter is going to fail in such a way that it may transmit incorrectly on frequencies other than is intended, and when we get reports from pilots or from our flight test folks of such occurrences, then we send out folks specially equipped to locate those ground stations and get them corrected.

MR. PHILLIPS: So, you're very dependent on the way the system is structured today to find the faults with the system?

THE WITNESS: That's correct. Changes in the electromagnetic environment, changes in the spectrum, changes in non-navigation systems on or near an airport are detected in general by the users. There's no present way to monitor throughout an approach, for example, the -- the cleanliness of the ILS spectrum.

MR. PHILLIPS: Does the Guam Airport area present any unique characteristics as far as ILS system approaches go?

THE WITNESS: Well, I think not. It's certainly got a lot of terrain, but we have many airports with terrain. We have -- when you have high terrain, you have hilltops and mountains which are very advantageous for other transmitting systems. People like to get their transmitters up at a high location.

So, it's fairly common that we will have AM and FM broadcast stations and various personal radio systems in and around airports and on high locations.

MR. PHILLIPS: There's a military base on the other end of the island at Guam, which operates an ILS system that's approximately aligned with the Runway 6 Left system at Agana.

Would you expect that to have any effect on the Agana, Guam, approach?

THE WITNESS: No. The -- the two ILSs that you speak of, the one at International and the one at the Air Force base, are, of course, on different channels because of the spectrum management activity that I spoke of earlier.

One of the components of assigning frequencies for ILSs is to assure that nearby ILSs are sufficiently apart on the radio dials, sufficiently apart in frequency, that common receivers can easily separate the two.

MR. PHILLIPS: Does the FAA maintain any kind of a database relative to interference or spurious signal cause and effect?

THE WITNESS: Yes. I'm a little hesitant about database. We have a logging system and a reporting system for interference cases, which may appear in some cases to look like a database, yes.

MR. PHILLIPS: Okay. Just a few closing comments here. I would be interested in your comments about future avionic systems designs relative to ILS systems, and in particular, the proliferation of electronic cockpit displays and the potential effects on the ILS systems navigation units.

Do you see a trend toward improving the margin of safety with the newer avionics versus the older designs?

THE WITNESS: Well, yes. As I mentioned earlier, it is increasingly easier and less expensive to produce better and better receivers. We've all seen how electronic systems continue to get cheaper in cost and generally have better and better performance.

So, receivers in general aboard aircraft are increasingly capable, and -- and now we are seeing a single box that has microwave landing system, instrument landing system, and global positioning system receivers all in the same space that a single receiver used to occupy.

Increasingly, with more and more software-based systems, the amount of hardware required is less. This means that the receiver itself has less complexity, less potential for failure and so on.

On the other hand, the software has the potential for failure, and, so, software quality assurance is becoming a very large component of receiver design.

The displays in aircraft are becoming more and more cathode ray tube and flat panel-based. These displays have a lot of electronics to drive them, and any electronics has a potential for generating signals. So, there's a corresponding increase in the amount of testing to ensure that on-board systems don't affect on-board receivers.

So, the standards bodies have been adding more and more tests for -- to ensure compliance that the signals emitted by circuits aboard the aircraft are not affecting aircraft receivers.

MR. PHILLIPS: And as a final question, are there active working groups in the aviation community looking at the issues of interference, spurious signals, and ILS system improvements?

THE WITNESS: Yes. Most aviation authorities have their own. For example, FAA has several, and I serve on a couple international committees which are editing and improving, updating ICAO and X-10, the document used worldwide for ground and airborne testing of nav aids and so on.

In general, to keep up with the changing environment that receivers operate in, higher and higher power broadcast stations and so on have resulted in a requirement, for example, starting very soon, that aircraft operating in international environments have to have a new receiver that's more immune to these off-channel signals.

MR. PHILLIPS: Do you expect in the future to see ILS systems replaced with another precision landing system?

THE WITNESS: Great question.

MR. PHILLIPS: My last one.

THE WITNESS: Certainly that is the general goal of most aviation authorities, is to migrate to satellite-based systems. However, there's a large portion of the avionics community that feels that at least as a back-up system, some small portion of the existing instrument landing system installation should be kept. So, I believe the technology will support moving to satellite systems.

MR. PHILLIPS: Thank you. That's all I have.

THE WITNESS: You're welcome.

CHAIRMAN FRANCIS: That's interesting. It's possible we'll get through the whole morning without talking about MLS.

I'd like to just make a comment and an observation here for those in the audience, and that is both at the NTSB and the FAA, we have what are called national resource specialists, and -- and these are people who, because of exceptional qualifications and international reputations, are designated to operate in certain areas.

It turns out that both Mr. Spohnheimer and Mr. Phillips are national resource specialists, and I think that the exchange that we've just witnessed is evidence of why they are. That really was extraordinarily interesting and informative.

Thanks to both of you.

I would now say to all of us here concerned that we would -- we would like to keep things moving along. So, let us all of us keep in mind that which has been said and try to avoid redundancy in our questions or going on longer than is necessary.

KCAB?

MR. LEE: Thank you, Chairman.

Mr. Phillips put special technical questions, and Mr. Spohnheimer gave us excellent answers, and I'd like to take this opportunity to appreciate both of you gentlemen.

Just one thing. Let me just double check. The KAL accident, the location was, as you know, --

CHAIRMAN FRANCIS: I thought he was so good that he'd be able to operate without one. Go ahead.

MR. LEE: The location of the KAL accident is Nimitt Hill, as you know. There are antennas and many other radio facilities located also in that area.

Given that, do you think in your personal view from the vantage point of a specialist, do you think all those radio facilities had any effect on the accident?

THE WITNESS: Statistically, I think it is unlikely, but it is very difficult to say with any certainty without some testing, and -- and even so, the nature of spurious signals and the failure modes that produce them means that as antenna systems change and deteriorate, the conditions change.

Certainly we have -- most airports are challenged with the same sorts of problems. I would offer in general that -- that I probably am aware of five or 10 cases in a given year of interference to an instrument landing system in the case of several hundred ILSs.

So, the occurrence is not rare, but it's perhaps in the one to five percent range.

MR. LEE: Thank you very much. That's all.

CHAIRMAN FRANCIS: Boeing Company?

MR. DARCEY: We have no questions, Mr. Chairman.

CHAIRMAN FRANCIS: Barton?

MR. EDWARD MONTGOMERY: No questions, Mr. Chairman.

CHAIRMAN FRANCIS: Korean Air?

CAPTAIN KIM: Yes, sir. We do have a question.

Not to delay the process, but would you please tell us if FAA ran any kind of testing on the localizer signal, interruptions or deviations, as well as the Korean Air-run glide scope testing, bench testing of similar nature to the localizers?

THE WITNESS: I'm not aware of any bench testing on localizer receivers associated with this accident. As I did mention, we -- we flight tested the localizer in the day or two following the accident.

CAPTAIN KIM: Right. The question is referring not to flight testing but bench testing with similar set-up to verify the results as Korean Air did.

THE WITNESS: No, I'm not aware of any testing. I would expect the results to be similar, however, that -- that one could inject signals that would cause the flag to move.

CAPTAIN KIM: Okay. Thank you very much.

May I ask you one more question? So, are there any plans underway to continue testing at Guam, in specific to find out if there are any more things to be discovered regarding this accident?

THE WITNESS: I'm not aware of explicit plans, but there has -- I have been a participant in some discussions about the nature of ways we might test the Guam environment more fully.

I did request an extra airborne test just recently to make some recordings of the ILS with the glide scope off the air. That was done within the past week. It took perhaps 45 minutes. So, it is only a very short look at the nature of the spectrum at Guam with the glide scope off the air. Nothing was found on that particular check, although it was a very short one.

CHAIRMAN FRANCIS: I think Mr. Phillips might supplement that answer.

MR. PHILLIPS: Yes. I'd like to comment on that. The systems group has had discussions concerning plans, potential plans for additional site testing at Guam in an attempt to identify potential signal sources.

One of the issues you may be aware of is that after the accident, there was a typhoon passed through the island that did considerable damage to the antennas and transmitting system there.

So, we believe that the environment at Guam today is different than at the time of the accident, but nevertheless we intend to -- to set up a plan to go take a look for -- for potential spurious signals. So, that's an activity that we'll be discussing in the systems group over the next couple of months.

CAPTAIN KIM: I'm sorry to delay, but we have one more question, and we have about 30 seconds before we ask this question, Mr. Chairman.

CHAIRMAN FRANCIS: Why don't we go to the other parties, and then we'll come back to you.

CAPTAIN KIM: I apologize. Thank you.

CHAIRMAN FRANCIS: NATCA?

MR. MOTE: Thank you, Mr. Chairman. Just a very brief question.

Sir, do you have any opinion as to the -- any particular technical difficulties, and just in very general terms, the cost of co-locating DME facilities with the ILS transmitters?

THE WITNESS: Yes. The cost of installation is quite minor, perhaps $10,000, if there's an existing building with enough room. The equipment, DME equipment, would be perhaps $100,000.

MR. MOTE: And are there any particular technical considerations regarding such an installation?

THE WITNESS: Well, there are many, but none particularly challenging. We have many installations with localizer and DME co-located.

MR. MOTE: Thank you very much, sir. No further questions.

CHAIRMAN FRANCIS: Steve, you ready?

CAPTAIN KIM: Yes, sir. We're prepared at this time. I understand there's not conclusive evidence to continue further testing of the equipment.

In particular to the Model 51RV-5B, are there any plans underway to improve the safety performance of this equipment in particular?

THE WITNESS: I'm not aware of any, but I haven't spoken to the manufacturer recently either.

CAPTAIN KIM: But nothing will be initiated from the FAA's part to mandate any kind of further improvements on that model?

THE WITNESS: I don't know how to answer that. I -- the avionics group, which happens to be located in Seattle but a different part of the agency than myself, would -- would have to initiate some dialogue to -- to promote such a change.

I take it you mean about the flag circuits?

CAPTAIN KIM: Yes, sir. You just described the process you would -- that's how you would go about it, but are there -- do you have specific plans at this point to initiate or mandate specific improvements to that model by the FAA?

THE WITNESS: I know of none.

CAPTAIN KIM: Thank you very much.

CHAIRMAN FRANCIS: Government of Guam?

MR. DERVISH: Thank you, Mr. Chairman. No questions.

CHAIRMAN FRANCIS: Mr. Donner?

MR. DONNER: No questions, Mr. Chairman.

CHAIRMAN FRANCIS: Mr. Feith?

MR. FEITH: No questions, sir.

CHAIRMAN FRANCIS: Mr. Montgomery?

MR. MONTY MONTGOMERY: Thank you, Mr. Chairman. I have just a couple short ones.

Mr. Spohnheimer, for the benefit of those of us who are not as technical as our national resource specialists, when we talk about injecting signals into the -- the device to see its response to a 120 hertz, I -- I -- I hear you say things like you just somehow squirt base band information in the system, and it responds in a way that's -- that's unsatisfactory.

However, in reading the report here, I find that they actually put this on top of an 88. -- 83.75 megahertz carrier in one instance and a 355 megahertz carrier in another instance. In other words, it takes a special way of doing this in order to get those base band frequencies in there, is that correct?

THE WITNESS: That's correct. The base band frequencies, the tones that we've been speaking about, are -- are called the modulation, and the carrier is the VHF or UHF signal that a pilot would tune a control head to.

So, in all cases, when I spoke of a signal generator, that was referring to a piece of laboratory equipment that could both generate the very high frequency signal, the numbers you're referring to, and add the tones to that signal.

MR. MONTY MONTGOMERY: So, if I walked up to this piece of equipment and played my radio real loud at a 120 hertz, it's not going to have any effect?

THE WITNESS: That's correct. I'm sorry if I left you that impression. The circuits that are sensitive to audio tones, of course, there's no microphone connected. They listen to those frequency-determining circuits that I had in the block diagram, and those circuits are the ones that limit the incoming signal to radio signals in the desired band.

MR. MONTY MONTGOMERY: And the modulation type, is it FM or AM?

THE WITNESS: This is amplitude modulation.

MR. MONTY MONTGOMERY: AM. So, if I flew over an FM station playing at 10 kazillion megawatts, what effect might that have?

THE WITNESS: Well, unfortunately, it probably would, even though the receiver is intended to respond only to amplitude modulation.

When -- when an FM signal is strong enough, it can actually affect the operation of the circuits and add amplitude modulation to the signal. So, a somewhat common occurrence among the interference cases is an aircraft operating close to a mountaintop-located FM transmitter.

So, the AM receiver is not immune to FM and vice versa.

MR. MONTY MONTGOMERY: Okay. Thank you very much. Thank you, Mr. Chairman.

MR. SCHLEEDE: Yes, Mr. Spohnheimer, I just wanted to follow up on one area you mentioned about the possibility of doing some testing at Guam to look for some problems.

Do you have any recommendations for us regarding our investigation whether we should be doing additional testing or the FAA should be doing additional testing at Guam?

THE WITNESS: Well, my view is that the FAA should be involved and perhaps has an incumbent responsibility to do something out of the ordinary to assure that there are no -- no extraneous signals affecting ILS.

The difficulty with all of this testing is that if -- if an extraneous signal is due to another user or to a degraded transmitter of some sort, they're seldom continuous. They're usually intermittent, and sometimes it takes a very long time to locate something that's clearly been reported.

It's like proving a negative. You can -- you can flight test it for a week or two weeks, and if you haven't found anything, you can't say that it didn't exist. Obviously if you find something right away, then you're done, and you can go fix it.

So, I think it would be reasonable to -- to define a -- a short test program that had a definite end to it that made a diligent effort to confirm that the spectrum is clear in the area of the approach.

MR. SCHLEEDE: And would the -- would the pilots that fly in there certainly play a factor in reporting outages or -- I'm sorry, not outages, spurious signals?

THE WITNESS: Yes, they would. The sort of thing that an engineer would probably do is -- is haul some test equipment out to the area and set it up with a computer so that it logs the conditions automatically every so many minutes for -- for some hours or days, so that we have actual measurements using lab-type equipment as opposed to user complaints.

But if there were approaches being flown at the time, it would be easy to add that sort of information certainly.

MR. SCHLEEDE: Thank you very much.

CHAIRMAN FRANCIS: Mr. Berman?

MR. BERMAN: No questions.

CHAIRMAN FRANCIS: Mr. Cariseo?

MR. CARISEO: No questions.

CHAIRMAN FRANCIS: Thank you very much, sir.

THE WITNESS: You're welcome.

CHAIRMAN FRANCIS: That was a very helpful and impressive performance, and those of us that travel a bit have -- are particularly impressed by anyone who travels 40 weeks of the year, and I won't ask you about your family status.

THE WITNESS: Thank you.

(Whereupon, the witness was excused.)

CHAIRMAN FRANCIS: We have five witnesses left. Three of the five, including Captain Woodburn, who is the next witness, are, I think, a little unusual for an NTSB hearing, but I thought that it was interesting to -- to perhaps have a little wider perspective on some of the issues that we consider important here.

So, Captain Woodburn is a captain with British Airways. He and I worked together on a number of committees, many of which are -- are involved with the CFIT issue, and I think that he can give us a contribution in terms of the overall worldwide implications of this kind of accident, and the same will apply to Don Bateman and to Jim Terpstra.

Mr. Schleede?

MR. SCHLEEDE: Thank you, sir.

Whereupon,

having been first duly sworn, was called as a witness herein and was examined and testified as follows:

TESTIMONY OF
CAPTAIN PAUL WOODBURN
BRITISH AIRWAYS
CHAIRMAN, ICAO, CFIT STEERING COMMITTEE
LONDON, ENGLAND

MR. SCHLEEDE: Captain Woodburn, please give us your full name and business address for our record?

THE WITNESS: It's Captain Paul Woodburn of British Airways PLC, The Compass Center, Heathrow Airport, London, England.

MR. SCHLEEDE: And would you please, because you're called here as an expert in this field, give us a summary of your experience and training and education that qualifies you for your current position and status?

THE WITNESS: Yes, sir, I will. I've been 34 years of flying with British Airways, 25 years as a captain and currently a captain on the Boeing 777.

I also have 23 years of flight management experience, the past 12 in senior management positions.

I also have 20 years of other industry experience having served on a number of industry committees and various projects. One in particular concerns this inquiry, and that is the Flight Safety Foundation initiative commenced in 1992 into Controlled Flight Into Terrain, CFIT.

I was a founding member of the original steering team. I have served as a member of the CFIT equipment team, and I'm now currently the chairman of the steering team for the past 18 months and a member of the implementation team for both CFIT and approach and landing accident reduction.

I'm also a Fellow of the Royal Aeronautical Society and a liveryman of the Guild of Air Pilots and Air Navigators.

MR. SCHLEEDE: Thank you very much.

Dr. Brenner and Captain Misencik will question.

DR. BRENNER: Mr. Chairman, we've asked Captain Woodburn to prepare a presentation about the industry efforts. With your permission, we'd like to have him present that.

CHAIRMAN FRANCIS: Go ahead.

THE WITNESS: Mr. Chairman, ladies and gentlemen, as you are expecting, I have a short presentation here to explain, I think, the problem of CFIT so that we can all understand it and, of course, to explain the Flight Safety Foundation initiative and to discuss some of the recommendations.

CHAIRMAN FRANCIS: Paul, could I just sort of reiterate the reminder that the interpreters are trying to follow you. So, I suspect that you're going to be more easily understood by them seeing as you're speaking real English, but -- but if you could sort of modulate your speed, I think they'd appreciate it.

THE WITNESS: Okay, Mr. Chairman.

I start with a definition of CFIT. There is no internationally-agreed definition, and the one on the screen in front of you reflects the one we chose for our work in the Flight Safety Foundation.

CFIT is when a perfectly-serviceable airplane is inadvertently flown into the terrain or water.

Can I have the next slide, please? Here you can see some statistics on controlled flight into terrain, and this reflects worldwide experience. On the bottom axis are years from 1968 through to 1997, and the vertical axis are the number of accidents predominantly to jet aircraft.

Over on the left-hand side, you can see where GPWS was introduced alongside the highest peak.

CHAIRMAN FRANCIS: Could we turn the lights down a little in here so we can perhaps get a little more better look at this presentation?

THE WITNESS: And you'll see the relatively dramatic reduction thereafter.

Over on the right half of this particular visual, you can see highlighted blocks, and I would draw your attention to the two peaks that stand up there, and they reflect the years of 1988 through to 1991, and then, of course, in 1992, there is a second peak, and this appears to be a regular characteristic of CFIT data.

There is a cyclical action here. Over three to four years, there is a rise to a peak, and then it diminishes. We don't necessarily know the answer for it, but we believe it's related to industry awareness.

When it reaches a peak, there is so much media attention and awareness that there is a natural, I think, reaction to it and therefore could explain the reduction.

It was that peak in 1992, the second of those two peaks, which led to the Flight Safety Foundation starting its initiative into CFIT and approach and landing accidents.

Next slide, please. Here you can see which sort of airplanes CFIT is attached, and you can see over on the bottom left side there, there are approximately five large commercial jet accidents on average per year worldwide, and this was the data that we had in 1992.

Interestingly, you can see the impact on large turbo-prop, regional commuter turbo-prop, business jet, and business turbo-prop aircraft, and over on the right-hand side there, the business turbo-prop have an average of 23 losses per year.

Next slide, please. This particular slide just shows from 1992 in top left there the initiation of the Flight Safety Foundation initiative. That led to a commitment and then formation of teams. I was involved from that very early stage, and then, of course, the teams worked for several years, and the final working group reports were delivered towards the end of 1995.

A further year was taken refining what we now know as the CFIT Education and Training Aid, and that became available towards the end of 1996, being distributed to industry in early 1997.

So, the bottom two lines there are concerned with Flight Safety Foundation implementation team activity which continues and the application of the associated products.

Next slide, please. Why did we concentrate our attention on CFIT particularly? This is worldwide and U.S. airline fatalities classified by type of accident over a 10-year period. The highest peak on the left-hand side there are the fatalities due to CFIT.

The next highest peak is the loss of control in flight, and that's another story, and it's because this particular peak of CFIT there that there has been so much industry activity.

Next slide. Where does CFIT occur? The simple answer is worldwide. This particular slide shows western-built commercial jet transports again up through to 1997. This is just a five-year period, and this is the latest data and not the data that we saw when we started our work. But let me talk you through this.

First and foremost, in the middle, in Eastern Europe and the Middle East, the figures of zero are not really zero. They are areas where we have insufficient data.

If I can turn now to North America, and you will see an accident rate there of .03. That has been pretty stable over a long period of time at that value. But over there adjacent to it, you can see Europe at .10. In other words, three times worse than North America.

Coming down to Latin America, you can see the figure there at 1.12, and that is a figure of 37 times worse than North America. These are CFIT accident rates.

Moving across to Africa alongside, that figure there is 18 times worse, and then moving across to Asia Pacific, these figures are 23 times worse or down in Oceana, 11 times worse than North America.

They sound terrible figures, but I have to say that since we started this initiative, the worst figures that we saw before were in Africa, which were 70 times worse. So, there has been a significant improvement from 70 times to 18 times, whereas in Latin America, there has been no improvement whatsoever. That figure is still 30 times -- 37 times as bad as North America. So, this gives you a measure of the size of the worldwide problem.

Next slide, please. If we can concentrate over on that right-hand bottom corner, where it gets into 1997, if we can just move it slightly, you'll see here the same block diagram that we looked before, but what I'd like to concentrate on is that for 1996 and '97, those black boxes, they are three CFIT accidents per year for '96 and '97. All of those black boxes were non-precision approaches. In other words, five out of six accidents in those two years were on non-precision approaches.

Indeed, the accident data shows that the risk on non-precision approaches is five times greater than for conducting precision approaches.

Next slide, please. Here, we're looking at commercial jet aircraft, again a 10-year period, from '88 to '97, and this is where these accidents occur on what type of approach, and there's 38 accidents here worldwide. The very large blue block there, which is roughly half of the cheese, half of the number, were on step-down approaches.

The interesting thing is most of those had DME available. There are only three accidents there which are over at the 8:00 to 9:00 position on precision approaches, and they relate to probable glide scope receiver failure, probable failure of a flight director to capture, and also a possible autopilot not being coupled. But they're a relatively small proportion of the whole, and the interesting thing is that 70 percent of CFIT accidents occur on final approach.

Non-precision approaches generally are much more complex than precision approaches. For many pilots, they are less familiar. They are more error-prone. They require more comprehensive briefing. They need particularly careful and accurate monitoring, and it is possible for complex step-down procedures for steps to be missed or to be taken out of step. In other words, to get one step ahead of the airplane could be fatal.

Such approaches also need much more carefully-managed airplane crew and checklist management, and it is a characteristic of many CFIT accidents that they occur when the crew is pre-occupied or distracted by other tasks.

Next slide, please. Where do they occur? As I mentioned, 70 percent on final approach, and that solid red line in the middle is where most of these accidents impact the ground. They're all in line with the runway, and, fortunately, as we shall see for the next slide, you can see here an idealized three-degree glide scope in orange, red, and then these are the flight paths of many accident aircraft underneath the three-degree glide scope. In other words, following paralleling a three-degree glide scope but impacting the ground on extended center line but short of the runway.

The parallel to Agana, Guam, is obvious.

Next slide, please. The Flight Safety Foundation overall goals were to reduce the CFIT accident rate by 50 percent in five years and that's this year.

The latest data that we have available shows that this goal has actually been achieved, albeit the data is still being assembled, and I have not got it to show you today.

The second goal here was much more challenging, and if you remember those worldwide accident rates I showed you, the worst in 1992 being 70 times worse than North America, under this basis, we would be looking for a rate no worse than twice North America.

So, we've made some improvement but certainly not to the extent of this particular goal that we set ourselves.

Next paragraph. So, who was involved on this industry participation? With the Flight Safety Foundation, we brought operators, manufacturers, to some extent regulatory authorities, although I have to say that the degree of participation by regulatory authorities has been disappointing. There was very little direct involvement in any of the working groups by any of the regulatory authorities worldwide.

However, they were kept informed of what we were doing either through the Flight Safety Foundation or by direct contact. Flight Safety Foundation also represents training organizations, and we had good participation there.

Wherever the Flight Safety Foundation found another initiative already going, we combined resources, and then everything was put under the Flight Safety Foundation banner, and that brought in ICAO, IATA, IFALPA, ALPA, the ATA, and again the ATC authorities.

Like the regulatory authorities, the ATC authorities were reluctant participants, too. The interesting thing for all of this industry activity, it's not just organizations but represents hundreds of individuals who have worked with us, some of whom still work with us on this particular initiative.

ICAO is normally recognized as a body that takes five to seven years to do anything, yet it has been remarkably supportive and productive to this process. Since 1994, there's a lot that they've done as we shall see.

Next slide. So, let me just recap on CFIT. It is this inadvertent flight into terrain or water. It does cause the greatest number of fatalities. The risks on non-precision approaches are greater, and they almost always involve the breakdown of crew coordination and monitoring.

Another factor which became very strongly evident in the analysis of all of this work was that there is no single measure that we can take to prevent CFIT. It needs a range of measures suited to a particular operator and the operating environment.

There is no new single piece of equipment that can be fitted to aircraft that will make CFIT go away. Yes, it may help, but in isolation, it is not the sole cure.

Remember also that any new equipment requirement takes many years to implement across the entire industry, and in many ways, it's the areas of the world that have the least problem that will fit the equipment first, and it's those other areas of the world where the greatest problem exists that will fit it last.

Industry must therefore take action now because we can't afford to let this risk go on unaddressed.

Next slide. This ICAO requirement becomes effective on January 1st, 1999, and if you remember the earlier slide in terms of small aircraft CFIT exposure, this was aimed at applying GPWS-fitted to the smaller airplanes.

You still have to remember that there are up to 200 heavy jet aircraft flying in the world today that have no GWPS-fitted at all, even after 20+ years of requirement.

Next slide. The GPWS warning functions described here are in effect the characteristics of a Mark-2 or subsequent model of GPWS, and the effect of this rather more stringent set of requirements for ICAO is that the early Mark-1 GPWS installations will need to be replaced by Mark-2 or better.

Next slide. There are a number of other changes that are being pursued in terms of instructions and training requirement for the avoidance of CFIT. There is also the requirement being framed for a company policy on the use of GPWS. Proposals in this direction being very detailed, which is why I'm not going through them here today, were being presented to the ICAO Council only last week. We await progress reports.

Next slide. This is a whole series of future ICAO actions, and I will only mention briefly some of the things associated with these headings.

Under the licensing and training, Annex 1, the proposed changes there are mainly to do with air traffic control language, skills and proficiency with requirements for improvement by 2001.

The next one down, charting, is concerned mainly with the adoption of colored terrain all minimum safe altitude contour presentation on charts to improve their readability and understanding by flight crew, particularly in the cockpit environment at night.

Operation of aircraft, the third bullet down there, there are a whole range of things, whether they be equipment and procedures, but typical things being discussed there are prohibition of the old altimeters, things like three-pointer designs or fixed drum-pointer design of altimeters which can easily be misread. There are still many in the industry in use today.

Under equipment, there is a requirement, an extension of requirement for ACAS, pressure altitude encoding transponders, forward-looking wind shear warning systems, and others.

Under procedures, there are new requirements and new emphasis on standard operating procedures, altitude awareness procedures, including the use of standard or automated call-outs, guidance on the use of autopilot, the incorporation of stabilized approach procedures concepts, etc.

The next one down, instrument approach procedure design, under PANS-OPS, there are particular changes there applicable to non-precision approaches concerning the optimum angle, and, of course, growing interest in the application of vertical navigation, VNAV, or FMC approaches.

Under air traffic services, there are new requirements regarding radar vectoring to avoid GPWS alerts as well as emphasizing and encouraging the implementation of MSAW of which we've heard a lot on this inquiry.

The last bullet there in terms of publishing a manual on CFIT avoidance is still under consideration. Further activity with ICAO concerns the translation of the Flight Safety Foundation education and training aid into ICAO languages, the other five beyond English. We're still awaiting a time scale for that availability.

Next slide, please. So, in summary, these are the ICAO sorts of changes. There is a need to train to ensure pilot response to CFIT ground proximity warning systems and so on.

Now there are two different ways of doing this. Many operators use a technique of during normal proficiency checks, inserting what some call an imaginary or glass mountain which generates a GPWS pull-up alert unexpectedly.

The problem with that is that the pilots may have been operating perfectly normally, safely, under their proficiency check, and they then have what is a rogue warning that seems to come at them with surprise. That can be considered negative training because it causes them to mistrust their basic normal procedures.

Another way of doing it is to still show how ground proximity warning systems work but in a more creative way. I'll describe a way that I know well particularly, and I know a number of other operators use it.

Modern simulator systems have good visual displays. When operating to an airport in the simulator database, under VFR good visual conditions, in mountainous terrain, it is very easy to take a vector that puts the airplane flying towards a potential conflict with the terrain. The briefing to the crew is let it happen, see what it looks like, and don't do anything until the ground proximity warning pull-up occurs.

The pilots are then left with this situation of watching the ground approaching, eventually filling the windshield in the visual display, and still the pull-up does not occur, remembering that 15 seconds or so to impact is typical of the characteristics of such systems. It could be less or marginally more.

So, when they get to the pull-up point, they're on the edge of their seats, can't stand the sight of it, and then, of course, pull up, they do the pull-up maneuver, and hopefully, if they've done the right technique, having watched the ground approaching, they will follow the required escape maneuver. They then have this visual image of what it looks like to be that close to terrain.

The next part of the exercise is to repeat it all in a different area, different bit of terrain, but they're now IMC. They don't see the terrain at all. For you pilots out there, I can guarantee you've got that visual image with you for several years after the event of having done that exercise, and when you fly in IMC to the pull-up point, you remember what it looked like visually. You don't waste any time. You get out there very quickly indeed, and it is an aggressive maneuver needed. Gentle ones or time taken to say is this real or false is not a luxury that we can afford.

Now that type of pilot teaching, I think, is very powerful and much more meaningful to them.

So, moving on to the second bullet there in terms of updating early ground proximity warning system installations, I've covered that in terms of Mark-1s being replaced by Mark-2 or better.

The third bullet is in terms of encouraging development and application of enhanced GPWS. We also need to provide precision approach glide scope guidance whether that comes from GPS, GNSS, RNAV, and so on.

I think we all recognize the need to eliminate the step-down non-precision approaches because the accident data says we should. We also need to encourage the expansion of approach radar coverage with MSAW on a worldwide basis, not just in the few countries that presently use it, and, of course, as we saw earlier, we're fostering the equipment of smaller transports with GPWS.

Now set against that, what actions have the regulatory authorities taken? Relatively little.

Now let's turn to the next slide, and here what I've tried to do, rather than go through a detailed presentation of all of the recommendations which would be beyond, I think, the scope of this inquiry, what I've tried to do is to show some of the applicable recommendations, and then I'll talk a little bit about them.

Chart supply and presentation. One of the recommendations was that looking at the worldwide data, a factor in some of the accidents was that not all crew members have charts. If they don't have charts, how can they effectively monitor what's going on?

So, there is a requirement that all crew members should have appropriate charts, and then, of course, the charts themselves in terms of presentation should have clear depiction of terrain and be easy to read in the cockpit environment. Hence the recommendation of colored contours.

The second bullet down in terms of approach and departure briefings, again the accident record shows that many of them have a failure to conduct adequate either departure or approach briefings. The more complex the approach, the more briefing and careful rehearsal of what is needed on that approach becomes necessary.

The third bullet down, allocation of flight crew duties and the use of the monitored or, as some call it, the shared approach procedure. An analysis of the accident data shows quite conclusively over hundreds of whole losses that they occur mainly in terms of IMC or at night, and on four out of five occasions, they occur when the handling pilot is the captain.

Another piece of data is that where crew coordination and monitoring is shown to be a causal factor for the accident, then it is four or five to one more likely to occur when it's the co-pilot monitoring the captain rather than the captain monitoring the co-pilot flying the approach.

That accident data therefore led to a recommendation that suggested that for IMC and night approaches, then the co-pilot should be flying the approach and the captain should be monitoring, and the captain takes over when visual reference has been achieved for the landing.

Now we all accept this question of the monitoring of the captain by co-pilots and so on is, I think, a worldwide cultural issue. The human factors experts have coined the phrase "the authority gradient". It applies to all nationalities, not just one particular nationality. All of us, I think, have a respect for rank, authority, experience, but in addition to that, there are some cultural issues, too.

It is more difficult for some cultures to be critical of the man in charge, the captain, or woman in charge than for some other cultures, and it doesn't matter how much training or whatever the company policies and procedures are, that has to be worked out continuously to achieve the correct, I think, crew integration and team effort. But all of that is part of this allocation of flight crew duties.

One other factor, a related recommendation, but I've not done it separately, is the use of the autopilot. Even for non-precision approaches, and probably particularly there, we've already discussed that it's a more difficult sort of approach. Why not use the autopilot? Because it reduces the workload. The handling pilot even operating the autopilot has more capacity to monitor what's going on and, I believe, will lead to a safer conclusion of that approach.

It does keep the workload well down, and I think improves this crew integration and monitoring enormously.

The last bullet is the non-precision approach procedures, including the design. This is where most CFIT accidents occur, and, of course, it led to the recommendations to try and make precision approaches more like -- or to make non-precision approaches more like precision approaches, where the accident rate is lower. It is the one most flight crews are performing most of the time. So, let's make non-precision approaches as similar as possible to precision approach.

The accident record of decades shows that jet aircraft have crashed for failure to follow stabilized approach concepts. So, let's incorporate stabilized approach into non-precision, which means continuous descent powers rather than step-down approaches which are inherently unstable.

It is also, as I mentioned earlier, very easy to get out of step with those -- those particular vertical descents flying level going to another one and so on.

There's a recommendation, too, that the construction of such approaches should be around the three-degree point provided all obstacle clearance can be achieved, and that one should have a final descent power of at least eight to 10 miles to allow stabilized conditions to be established more easily than trying to do it from the final approach fix at four miles in-bound.

I'd like to just look at a chart at the moment, and this just shows a particular instrument approach chart. It's an ILS or a VOR to Runway 8 in Gabaroon or in Africa, and this is a particular approach with similar characteristics to Guam. This has a VOR DME at the final approach fix.

Down there in the bottom right-hand corner, if we can zoom in, bottom right, that's it, just there, you'll see DME distance with an altitude table, and this is the sort of information that a company I know which produces its own charts provides to pilots which gives additional DME guidance beyond that final approach fix to determine an optimum descent angle, and that actually computes to a 3.1 degree angle.

If we now go back to the profile, and we can see there in the middle the GBV VOR DME at the final approach fix, crossing altitude of 4,800 feet, there's nothing whatever to stop anybody commencing a final descent path instead of 5,300 shown some distance out. It only needs to be less than two miles outside that VOR DME to do a continuous descent path.

Indeed, you could run right around the whole procedure at 5,500 and face finals at 5,500 and commence descent at 2.5 miles before GBV and do a continuous descent all the way in. You've observed all of the limitations, but you have a more effective continuous descent and stabilized approach capability.

Now, this chart is not ideal, but that's the sort of thing that I would like to see us eventually rewrite such procedures using the aids available and to allow the pilot to operate the airplane in the best possible way.

If we now just look at the planned view of the chart itself, that's the upper half, again all I would just draw attention to are the colored areas there in light green, and the figures in there. These are minimum safe altitude contours. In other words, the figure you see there is a safe altitude to fly at.

Now that's one way of depicting contour presentation rather than the terrain itself. This is, after all, what the pilot wants to know. What's the safe altitude I can fly at? Not necessarily read the height of the ground, apply a margin, and then eventually get to the figure. This is prime presentation of information.

Next slide, please. Coming back to these applicable recommendations from the CFIT education and training aid, the next bullet here is altitude awareness, and here, it's important that the flight crew establish the applicable minimum safe altitude for where the airplane is going to be and where it is.

They also have to bear in mind that the minimum operating altitudes, when in low temperature or high winds, needs to be increased, and that, I think, is a correction that is not well understood worldwide for international operators who may occasionally operate to either very low temperature airfields or indeed may experience high winds when operating at low altitude.

Altitude awareness also includes the incorporation of the 500-foot radio altitude call-out, particularly on non-precision approaches. The value of such a call-out, if integrated into normal operations, is that it's in the vicinity of most minimum descent altitudes.

When 500 radio goes off, if you're not close to being visual with the runway, then you should be getting out of there. That's the intent of that particular call-out.

There's also a requirement here that there is rather more positive cross-check of the final approach fix crossing altitude before continuing the descent to the runway.

The next bullet is radio altimetry and call-outs. It is vital, the accident record shows, that we have improved terrain awareness. Most of our aircraft have the radio altimeter on board our aircraft, but many operators don't use it for normal operations and only require its use in Category 2 or 3 conditions.

The significance of that is that in Category 1 or even in VFR conditions, then one should have the radio altimeter as part of the instrument scan when below 2,500 feet and lower commencing the approach. The intent of it is to make pilots aware that they are getting close to terrain and need to be aware of it.

Another feature is that how do you integrate it? Do you have manual pilot call-outs or, better still, have automated call-outs through the ground proximity warning computer? That has a number of menus of call-outs, and many aircraft have them today. The value of automated call-outs is that it doesn't get tired, distracted or anything else. When crews can forget to make the manual call-out, the automation doesn't. But the important thing is to have procedures associated with it, not just to have the call-outs made and then ignored.

The next one down here is measurement and evaluation of system performance. The world's airlines have imperfect systems quite often to measure how their aircraft are being flown, whether the standard policies, procedures and so on are being observed, and to what standard.

Here, what I am recommending and what the Flight Safety Foundation recommends here is the adoption of flight operations quality assurance sorts of programs, the foci as we know in North America, and comparable programs elsewhere.

A growing number of airlines are now using such data which means analysis of either flight data recorders, quick access recorders, enhanced pilot reporting, whatever, to monitor how the aircraft are being flown, and that information can be used for routine engineering purposes or operational purposes.

Sticking with the latter, it is possible to determine if limitations have been exceeded, flat-limiting speeds, for instance, whether the aircraft had a rushed approach. In other words, mismanaged approach by the flight crew.

Another one is a recording of ground proximity warning system alerts. There is too little data being collected by most operators as to how ground proximity warning systems are working on their aircraft.

I think many of us know that accidents show that pilots were ignoring the ground proximity warning shouting at them when the accident occurred, and they were ignoring it.

The big question behind that is why? Now, we know that the false or nuisance activation of ground proximity for some systems can be high, but there are technical solutions to make them much more dependable, and therefore pilots should be encouraged to believe them.

But getting the data is half the problem. When you know the problem, you can then apply solutions. You can also see how flight crew responded to GPWS alerts. That, of course, has benefits in terms of having confidence that this safety system is protecting your aircraft, but also the technique that is being applied by the pilot in the recovery maneuver. Again, that can be fed back into the training program to refine the technique, and then by gathering the data after the change, you can measure the improvement.

Another one is monitoring of go-arounds. We've talked a little bit about that earlier in this inquiry, and we believe it's important to monitor for a variety of reasons. Yes, we must not discourage pilots to perform go-arounds when necessary. Indeed, you must positively encourage them.

However, you need the data for these sorts of reasons. For instance, at congested airports these days, aircraft are being squeezed in to maximize capacity with minimum separation between airplanes. It is possible that by monitoring go-around rates, you may find a problem at one particular airfield. That may need a discussion with air traffic control to refine their procedures.

Another benefit could be not just the numbers of go-arounds but how are they performed. We all know that in the simulator on proficiency checks, pilots perform the required maneuvers well. They have to. They're being assessed on it, and nobody worries about doing aggressive go-around maneuvers in simulators.

However, most pilots change when they've got 400 passengers sitting behind them on an aircraft, and there is almost an unconscious relaxation, an attempt to be somewhat smoother, gentler. The reality is you do a lazy go-around by comparison with an aggressive go-around, and when you are in the vicinity of low minimum descent altitudes or decision heights, getting close to the ground, you cannot afford that luxury.

So, again, if you monitor performance, you can feed this thing, feed the information back into the training and the education of your pilots.

The last one on here is the minimum safe altitude warning system. I won't go into any more detail on this, but there are recommendations about its worldwide application. It is available in many countries, as we've already heard, but it is in limited use worldwide. We need to see more of it.

You can take the slides off now, please. In conclusion, I don't have a slide for this but would just like to make a few remarks.

In spite of the efforts of the Flight Safety Foundation, the many individuals, some of whom are in this room today, and in spite of what we've now discovered about controlled flight into terrain accidents worldwide, they still continue to occur. Just ponder that. They still occur.

I believe that industry needs some degree of compulsion to take more effective action. It's not enough at the moment to have awareness and voluntary action. We need the help and support from the regulatory authorities to maintain the momentum of this Flight Safety Foundation initiative and the work that the industry has completed.

Remember the ICAO proposals that are being worked on now need state approval. State authorities will listen to their regulatory authorities. So, we need the support from the regulatory authority to ensure success of those ICAO proposals.

But that's not all. I believe all public transport operators should be required to have a CFIT avoidance strategy and a program with policies and procedures applicable to that particular operator and its operating environment, but based upon the Flight Safety Foundation education and training aid.

It's then not enough to have policies and procedures. The regulatory authorities must verify that they are in place and being used.

Operator training programs should incorporate the diverse nature and range of instrument approaches that they encounter in the real world in their simulators.

We should also recognize that continued development and application of new technology and equipment, both in the air and on the ground, should be positively encouraged.

Chairman, ladies and gentlemen, thank you for this presentation.

CHAIRMAN FRANCIS: Thank you very much, Paul.

I'd like to preempt perhaps a little bit a question, but -- but I think that this issue that -- that you mentioned several times of participation in the groups that are working on this and particularly participation by regulatory and air traffic authorities is extraordinarily important.

We basically have a situation where as far as I can see, the entire rest of the industry is involved, and yet the people who are essential to -- to moving much of the -- of the equation here that we're talking about are not involved, and I'd be interested in any thoughts you might have as to, Number 1, well, particularly why they may not be involved, and I certainly hope that this hearing and anything that we can do afterward to get them involved, we can all work on.

So, if you have any comments on this. We didn't co-conspire, by the way, on this, but -- but I think we're both coming from the same place.

THE WITNESS: Well, thank you, Chairman. Yes, it has been a difficult area. I think one recognizes that not just operators but regulators have also had difficulties with resources and processes of change and various other internal problems.

It has been difficult for them to resource these sorts of industry activities, but the converse of that is that we found it difficult, too, but felt it important enough to do it.

That, I think, is the -- the message that now needs to get to the regulators, that the work and the progress that has been made will not be maintained unless they join this program.

I know my own regulatory authority in the U.K. I have given presentations to them on this, and they have been reluctant to take it on as a regulatory activity.

Remember when I suggested that some encouragement be given to it. That's one thing, but verification means more work, and that maybe is what they're hesitating over. But I don't think we have the choice. The data shows that this is the biggest cause of fatalities, and we must react to it.

It would be a very powerful, I think, signal to the world if we could persuade, for instance, the FAA, either as a recommendation of this inquiry or beyond it, to come on site and to take a more active role in running with the recommendations that have come out of the Flight Safety Foundation.

There are no axes to grind here. We have a shared common goal, safety.

CHAIRMAN FRANCIS: My apologies to the Tech Panel for that, but I think that's an extraordinarily- important message for us to get across, and proceed.

DR. BRENNER: Thank you, Mr. Chairman.

You mentioned that there's been a -- a major reduction in CFIT accidents since the beginning of this effort. What -- what are some factors, do you think, in helping in that reduction so far?

THE WITNESS: I think the major factor has been the increased awareness within the industry, and certainly since the Flight Safety Foundation commenced this initiative, there has been a lot more media coverage of this activity.

The combination of this, I think, awareness and the growing availability now of products like the CFIT checklist, the various videos in both corporate aviation and that comes with the education and training aid, these are the sorts of things that are now being more widely applied within operators.

But I believe a great deal more needs to be done to maintain the limited improvement that we've seen thus far. We'd dearly like to see this problem eliminated.

DR. BRENNER: You mentioned the checklists, the CFIT checklist. How is that used?

THE WITNESS: The CFIT checklist, for those of you that may have seen it, is a fairly complicated list of factors which enables airline management, not operating flight crew members, but airline managements to assess the nature of their operation and to come out with a risk-degree factor at the end of it which may cause them to select appropriate measures that reduce that risk, various policies.

I mean, for instance, how one flies non-precision approaches or the use of the monitored or shared approach, those sorts of things. They are mitigating factors against a risk of a particular type.

So, yes, it's a management tool, not an operational tool.

DR. BRENNER: Would -- would pilots use it as well?

THE WITNESS: I don't believe they would find it very user-friendly, no. I think most pilots want things much shorter, sharper, punchier, whatever, and we already have difficulty with long checklists in airplanes now.

The CFIT checklist is quite complex and really is not a factor for them because most flight crew are not the determinants of operating policies and procedures. That's the airline management's.

CHAIRMAN FRANCIS: Excuse me, Malcolm. This is aimed -- the CFIT checklist is aimed at the issue that we're all talking more and more about, and that is that safety is not just the pilot ran the airplane into the water or into the mountain. Safety is ultimately the responsibility of corporate management in whatever company it is, and that this starts at the top of the management.

So, this checklist, while complex, is aimed at what the entire spectrum of the company is doing in terms of its policies in order to prevent CFIT. It's aimed at the company and not just at the operations people, but in the -- at the entire company.

THE WITNESS: If I might diverge very slightly, Chairman, there is some work going on in another country, which is trying to enhance what I would call awareness of safety management systems, and there, they have already discovered that the most important factor on the safety performance of any organization is its management culture.

Have the right management culture, safety in terms of both culture and performance will result. So, it's just really emphasizing the point that you made that safety starts from the top, doesn't stop there. It runs right down through the organization from top to bottom and all the way back up again. It has to be, you know, staffed. It has to be resourced. It has to have an organizational commitment to safety in everything that that management organization does.

DR. BRENNER: Captain, in the case of the accident flight, would the checklist have highlighted certain areas of risk that might have developed more attention?

THE WITNESS: I believe that the use of the checklist will highlight to management, yes, that certain types of operation do have higher risks and that there are policies and procedures that could reduce that risk when applied. But as I say, it is for managements and not the operating crew.

DR. BRENNER: The -- you mentioned that the CFIT training aid was sent out last year. How has the response been from the international community?

THE WITNESS: That is an interesting subject. We know that more than 2,000 copies of the education and training aid have been distributed worldwide through the manufacturers principally and through some training organizations and other industry bodies.

The difficult thing is we now have to gather data as to what airlines have done with it. We have no established communication at the present to measure that implementation progress.

So, the Flight Safety Foundation is considering sending out some form of small questionnaire, quite deliberately not aimed at where the CFIT education and training aid was sent. If it went to the VP, Flight Operations, and he did nothing with it, then it's no good sending the questionnaire to that particular individual.

What we'd like to do is to send the questionnaire to some lower point in the organization, for instance, into the training management arena, and also to the flight crew community themselves through the pilot associations.

We then have a measure of how effective changes might have been within the organization and the degree of communication on CFIT that's going on from top to bottom.

Now that data-gathering is due to commence later this year, and we will be eventually reporting on what we find to the Flight Safety Foundation, and we hope that that information can be used to both encourage the airlines that have started doing something and, I hope, to prompt those airlines that have done very little so far to start doing something quickly.

DR. BRENNER: How many airlines are using monitored approaches?

THE WITNESS: I don't have an exact number. All I can say is that there are a large number and a growing number now using the monitored approach, if not for all of their operations, at least for part of their operations.

The name of it may vary from one airline to another. I've already used the term "shared approach". Some airlines use the term "low-visibility procedures approach". So, they may have a different set of procedures for Category 2 and 3 that may be different to the procedures used for Category 1 or VFR flying.

There are also a number of military forces in the world that use it, too, particularly in the transport arena. So, yes, it's being more recognized and steadily growing.

DR. BRENNER: Among airlines that have hesitated to use this approach or decided not to or are considering it, what are some of the concerns that are raised?

THE WITNESS: There is a difficulty when an airline has an established operation that may have existed for many years, and pilots are resistant to change. It's remarkable how pilots can adapt to new concepts with a new airplane that they're required to fly but are remarkably resistant to changes of policies and procedures because they defend that which they know best.

So, airline managements who wish to make a change have a fairly uphill education task as well as a redefinition of policies and procedures to support the change.

It then doesn't happen overnight. I know from my own personal experience that it can take many years before these sorts of changes of concept can be fully accepted. But you only have to look at that accident data, and it's difficult to refute it.

There is a better way of flying airplanes. We know that. The data supports it.

DR. BRENNER: How many airlines have training for aggressive response to a GPWS warning?

THE WITNESS: Well, all airlines would claim to have it. I think only those airlines that have some form of system to measure performance in the way I was describing earlier know whether their pilots are actually doing it.

Simulator performance is not enough. You have to see what they're doing on the real airplane. I don't have figures of how many airlines are doing aggressive. I just know that that is the general policy, but few airlines have the means to ensure that it's being done.

DR. BRENNER: Yesterday, we spoke about considerations of tracking missed approach data. Do you have any -- any thoughts on that, on any value towards this type of effort?

THE WITNESS: I'm sorry. Could you redefine that question a little?

DR. BRENNER: I believe keeping airline records on go-arounds.

THE WITNESS: Oh, yes. We -- we keep the records. We feed the information back, and I think operators generally in being encouraged to keep the record should do that.

As I indicated, it does identify problem airfields with other causes for go-arounds, but the important thing is that we use it for beneficial purposes in terms of encouragement and also the correct performance of the go-around itself.

It is essential that the aggressive maneuver for a go-around is performed when at or near the minimum descent altitude or decision height, but when you are well away from it and commencing a go-around from more than a thousand feet away from such low altitudes, could be more gentle, and that may be an airline policy choice, but again have the data, use it, refine it, and then have confidence in how your pilots will perform.

DR. BRENNER: And we spoke yesterday about MSAW. Are there international standards or requirements?

THE WITNESS: There are none yet, and that is the work I referred to earlier in terms of ICAO. Have a proposal to mandate it at some point in the future.

However, we know that for recent radar equipment installed worldwide, most of them have the MSAW capability. Other than a few states, like North America, like Israel or Turkey or one or two other places in the world, most do not have them commissioned. They do not have them tailored to the installation.

Air traffic controllers are not trained in its use, and indeed there is some degree of air traffic resistance because, remember, MSAW, an alert, could indicate that the air traffic controller made a mistake, and there are some therefore cultural or punishment issues associated with that alert, which are natural inhibitors to adoption.

But all of those issues have to be worked through to make sure that we do have the safety benefit that is available but being unused. In other words, the cost of actually putting it in place is minimal.

Let's use it.

DR. BRENNER: Is there CFIT prevention training for air traffic controllers?

THE WITNESS: There isn't, but there should be, and that was one of the recommendations that came out of the air traffic control procedures and ground equipment working group report, and the sort of things that need to be done are training to understand the capabilities and requirements of aircraft.

I think many of us take that for granted, but I believe air traffic controllers need to have more knowledge in that area. They need to understand the stabilized approach procedure and what it means to us as pilots when they ask us to fly at certain speeds to certain short distances from touchdown.

They need to improve their awareness of GPWS performance and radar vectoring in the vicinity of terrain. They also, I think, need to have education in terms of operation at low temperatures or high winds when operating at low altitude in the vicinity of terrain.

Many states have no procedures for such -- for such conditions. Others have procedures where air traffic control will modify clearances. Other states have procedures where they expect pilots to make the corrections and then notify air traffic of such corrections. There is no uniform standard, but there should be.

Those are the sorts of areas that I would see education needed.

DR. BRENNER: The NTSB has recommended to the FAA to make CFIT training mandatory for airline pilots, like wind shear -- training in wind shear avoidance. Is this a positive step?

THE WITNESS: That's a positive step, but as we have seen, and I -- as I have tried to reiterate, there is no single step that stops CFIT. It is a collection of measures.

CFIT education and mandating of it is just one element of those measures. Another piece of equipment on the airplane is not the only measure needed. It is a step in the right direction.

DR. BRENNER: Are there some measures that can be implemented immediately?

THE WITNESS: Well, interestingly, most of those things I talked about in terms of applicable areas of the Flight Safety Foundation education and training aid report, most of those areas could be applied at little or no cost.

What it requires is management will to do it, and then, of course, a resource and effort to support it. So, there is a small cost, but it's not a big one. We've already covered, I think, the crew education and awareness as being one step, but the most important thing is to make better use of the available equipment that we have on our aircraft. Some operators do that already, but many could make better use.

There needs to be a management review of policies and procedures. That takes time and effort, but it's well worth it. There needs to be appropriate and more effective training.

We also need to encourage, I think, the new equipment development and the application of new technology, and most important of all, we need to move in this area of performance monitoring so that we know how the aircraft and how the flight crew are performing when they're out in the airplane, not just in the simulator.

DR. BRENNER: Thank you, Captain Woodburn. That completes our questioning, Mr. Chairman.

CHAIRMAN FRANCIS: FAA?

MR. DONNER: Thank you, Mr. Chairman. We have no questions.

CHAIRMAN FRANCIS: NATCA?

MR. MOTE: Thank you, Mr. Chairman. No questions.

CHAIRMAN FRANCIS: Guam?

MR. DERVISH: Thank you. No questions.

CHAIRMAN FRANCIS: Korean Air?

CAPTAIN KIM: Thank you. No questions.

CHAIRMAN FRANCIS: Branson? Barton. I'm sorry.

MR. EDWARD MONTGOMERY: Thank you, Mr. Chairman. No questions.

CHAIRMAN FRANCIS: Boeing Company?

MR. DARCEY: Thank you, Mr. Chairman. No questions.

CHAIRMAN FRANCIS: KCAB?

MR. LEE: No questions. Thank you, Chairman.

CHAIRMAN FRANCIS: Mr. Feith?

MR. FEITH: Just several questions, follow-up questions, and the first one is probably tell-tale on ourselves.

You had spoken of the reluctance of regulatory authorities to become involved in this -- in this program, and you had spoken specifically of your regulatory authority.

Have you had any feedback as to the reluctance or a perceived reluctance on the part of the FAA or any other worldwide regulatory authority what their concerns are?

THE WITNESS: I've had no specific feedback to me personally at all. I have good contacts with my own regulatory authority, and they in principle support what's going on.

The problem is manpower to commit to doing it, bearing in mind all of the other tasks that they're supposed to be doing. That, I think, is more the heart of the problem, not an objection in principle, to what we're trying to achieve here, and it's a question, I think, of just changing priorities and recognizing that this is a valuable initiative that must be supported and continued to achieve the desired improvement.

MR. FEITH: I'll take that one step further and go beyond the regulatory authorities. I may be telling tale on ourselves, but has the NTSB or the AAIB or any other safety organization around the world been involved in this program?

THE WITNESS: Yes. I have to say that whereas CFIT may not have been supported as well as we would have liked, what I didn't describe to everyone here today was that the Flight Safety Foundation initiative concentrated on CFIT initially because of the fatality data that I showed you.

We also recognized that CFIT and approach and landing accidents are very closely related. Indeed, it's sometimes difficult to separate the two. It's really two sides of the same coin in some respects.

There are a number of working groups still running with the Flight Safety Foundation on the approach and landing accident reduction element of this initiative, and that has now involved both regulatory authorities and safety organizations, and I have to say I think that is after the event and the degree of success that CFIT activity showed. So, yes, we've got them involved at last.

MR. FEITH: And just to make sure that I have a correct perception, were the ATC authorities involved in the -- in this program, also?

THE WITNESS: They were invited to participate, and indeed the air traffic control and ground equipment working groups started off under an FAA chairman several years ago, but within a year, he took early retirement, and that was the end of FAA participation of any sort, unfortunately.

Subsequently, when the group was reconvened and then completed its report some 18 months or so ago, there were few representatives, if any representatives, from air traffic managements, but we actually had air traffic controller participation. So, it was the -- like the pilot, we had the man on the spot there.

MR. FEITH: So, that's worldwide air traffic control, --

THE WITNESS: Yes.

MR. FEITH: -- not just the FAA or --

THE WITNESS: Yes.

MR. FEITH: -- just --

THE WITNESS: That's correct.

MR. FEITH: -- that kind of organization?

THE WITNESS: Yes.

MR. FEITH: Just for a clarification, you had spoken in one of your presentations about stabilized approach velocity. I think just for the benefit, could you give us the nutshell or Reader's Digest version of what you mean by stabilized approach criteria for the approach segment because I think you related it to the three-degree approach?

THE WITNESS: I don't have the benefit of a diagram here, but if we can visualize a final approach segment around the three-degree descent path, and ideally one should have somewhere between eight to 10 miles of in-line approach, constant descent from what may be 2 to 3,000 feet, in a landing configuration established early enough such that the landing check-list can be completed and out of the way, to allow the flight crew to then perform the remainder of the final approach and the transition of the final approach fix without having distracting and conflicting tasks.

You then need to set gates at various points on the approach, and many operators choose, for instance, 1,000 feet above the field as a particular point when the airplane must be in the landing configuration, must be at the right speed at no more than maybe 20 knots past the target speed with the approach pass and landing checklist complete and so on, and that is the target for all approaches.

Then operators may have another point down the approach, and 500 feet is common, at which point there is a tighter gate still in terms of speed and associated conditions being on the vertical profile in the right position to complete the landing, and if the tighter set of conditions are not met, then there should be a mandatory go-around requirement from the 500-foot point.

The target at 1,000 feet, if not met, is one which has consideration given to go-around, yes or no; 500 feet mandatory go-around if the conditions are not met, and the final check is at 100 feet, and particularly on limiting runways, this target is where the aircraft has to be at the right point above the threshold, at the right rate of descent, and not exceeding a speed of, say, 15 or 20 as the maximum condition for landing.

On a limiting runway, if that particular gate is not met, then again mandatory go-around. So, I would liken it a bit like if you can imagine at 1,000, 500 and at 100 feet, three eyes of a needle. It's a slightly bigger hole at 1,000, a smaller home at 500 feet, and a very small hole at 100 feet, and you thread the aircraft through the needles, and you get it right.

MR. FEITH: Thank you, Captain, for that explanation. Appreciate it.

You had made a statement regarding providing all crew members with charts, so that they could basically all be up to speed on the approach. Would that include non-flying crew members; that is FEs or international relief pilots that may be in the cockpit but not actually performing a flying duty?

THE WITNESS: You added a caveat on the end there, not performing a flying duty. There are some two-crew aircraft designed for two-crew operations which have third crew members which do not have assigned duties, and that is one category, and I would say in that case, it's the operating crew members that have to have the charts.

However, there are many three-crew aircraft operating in the world today with either flight engineers as pilot or engineer in the third seat who are forward-facing for take-off and landing and who do have assigned duties of monitoring the pilots.

If they are to monitor effectively, they have to have the chart to be able to do that. It's very difficult in night-time conditions to be looking over a pilot's shoulder trying to reach his chart when you're supposed to be doing other things as well, or during briefing to try and extract and write down relevant bits of information to enable the monitoring to take place. That -- that procedure, I think, is unsatisfactory.

MR. FEITH: And with regard to one of the charts that you showed depicting minimum safe altitudes and your explanation that pilots would rather see what the minimum safe altitude is than to try to figure it out, ball park it and then make sure that they hit the right altitude, the chart that you showed is produced by an independent organization over in your side of the world. Jeppesen, of course, is typically a world standard for charting.

Do you have any comparison because Jeppesen doesn't show that on their charts? Do you have any particular opinion about the differences in charting?

THE WITNESS: Yes, I do. I mean I'm not being critical of any particular company. I believe that the industry recognizes that terrain or minimum safe altitude are better presented in contours rather than in tabular or spot height form.

One gets a much better impression. I actually have two charts here to show a comparison of the two different techniques which I could show, if you would allow me.

MR. FEITH: Please.

THE WITNESS: And you can see therefore the difference of presentation that one should, I -- I emphasize, not be critical of either. They are satisfying two different purposes, and the rationale behind it is to make it easy to read and use.

Now both are better than the ways that used to be the norm, and I would encourage developments in this direction. The problem, of course, with minimum safe altitude compared with presentation of terrain is that you may need some degree of skill and cartographic application to, as it were, draw the right minimum safe altitudes versus terrain which is fixed to the ground or topographical charts. You may simplify those, but they're easier to draw.

So, let me just show you, and you can see what they look like.

MR. FEITH: And just for the benefit of us, we're going to give Jim Terpstra an opportunity to defend his position when -- when he testifies regarding Jep charting.

THE WITNESS: What you've got here are two real charts.

MR. FEITH: Excuse me one second. Can we just lower the lights a little bit so we get a better picture, please?

THE WITNESS: Is it possible to focus that slightly differently? Okay.

Here on the left-hand chart, this is minimum safe altitude presentation of the safe altitudes to fly at, and you can see, I think, pretty quickly that it's very easy to pick the appropriate figures here. They're in hundreds. The large digit is the thousands, the smaller digit being hundreds, and it's very easy to then -- this -- these are the mountains to the southeast of Geneva.

If we look at the -- the chart on the right-hand side, here we're seeing -- if we go to the same area, the bottom right, one has to be a little bit more careful in terms of reading the figures on here and remember this is terrain. So, you've got to get the right figure, then apply the right margin of either 1,000 or 2,000 feet obstacle clearance, and remember you've got to do this in night-time cockpit conditions with the airplane flying at various speeds.

Now both of these presentations, this one is in a brown tint which shows the ground, the other chart is showing minimum safe altitude, which is in green, these conform with the ICAO Annex 4 requirements for charting, which says that the ground shall be in either black or brown, and that minimum safe altitude shall be in green.

Now, obviously one can see the basic similarity of terrain is evident on both, but you have to ask yourselves which is the easier one to use and apply flying an instrument approach.

We, in my particular company, started off using the terrain contour presentation some 35 years ago, and we then found some difficulties of interpretation in the night-time configurations of those aircraft.

We started to experiment with this type of minimum safe altitude display, and in the 1960s ran a test with our pilots, and we had a more than 90 percent in favor of presentation of minimum safe altitude rather than the terrain itself, and for the past 30 years or so, we have maintained this style of presentation.

Either of these, as I continue to reiterate, is much better than those earlier charts which did not have the contour presentation on at all. So, the fact that the industry is now moving in this direction is, I think, enormously important.

MR. FEITH: Thank you, Captain.

Lights up, please. One last question. You had talked about trying to collect, I guess, real world data from line operations, so that you cou