The U.S. Air Traffic Control System Wrestles with the Influence of ACAS

28 Pages·1999·0.41 MB·English

by  Flight Safety Foundation

The U.S. Air Traffic Control System
Wrestles with the Influence of TCAS
Although TCAS II has been credited with averting inflight
aircraft collisions, its implementation continues
to cause disagreements between pilots and controllers.
by
V.J. Mellone and S.M. Frank
A significant increase in incident reporting related pilot-controller interactions and to suggest strat-
to Traffic Alert and Collision-avoidance System egies that could promote a more harmonious
(TCAS II) implementation prompted a closer ex- human-technology interaction.
amination of how TCAS II is affecting pilots and
air traffic controllers. The study was limited to a randomly selected data
set of 174 ASRS reports from the January 1992-
In July 1992, the U.S. Federal Aviation Administra- March 1993 period. The data set specifically in-
tion (FAA) and the U.S. National Transportation cluded pilot and controller reports that involved
Safety Board (NTSB) asked the U.S. National Aero- traffic advisory (TA) and resolution advisory (RA)
nautics and Space Administration’s (NASA) Avia- incidents.
tion Safety Reporting System (ASRS) to complete a
data base analysis of TCAS II incident reports. A TA identifies nearby traffic meeting “certain
minimum separation criteria,” according to an FAA
The ASRS research team identified evidence of advisory circular (AC) issued in August 1993. An
increasing air traffic controller consternation with RA provides flight crews with aural and display
and resistance to the implementation of TCAS II information advising whether or not a particular
technology into the U.S. airspace system. There maneuver should be performed to maintain safe
were also strong indications from the data set that separation from a threat aircraft.
the aviation community, government agencies and
industry may have unwittingly underestimated the The research addressed four issue areas: Were there
influence of TCAS II avoidance maneuvers on air any significant changes in TCAS II reporting dur-
traffic controllers and flight crews. ing the past three years? Do TCAS II avoidance
actions influence the air traffic control (ATC) sys-
The study’s objectives were to analyze the ef- tem? Is there evidence of contention between pi-
fects of TCAS II avoidance applications on lots and controllers because of TCAS II applications?
FLIGHT SAFETY FOUNDATION • FLIGHT SAFETY DIGEST • NOVEMBER 1993 1
Are there effective strategies that could enhance system. In 1987, the U.S. Congress mandated the
pilot-controller cooperation? installation of TCAS II on all air carriers with 30
or more passenger seats by the end of 1993. To
The ASRS reports were randomly drawn from a date, more than 4,000 air carrier and business air-
pool of 780 TCAS II incidents from the same pe- craft are TCAS II equipped.
riod involving TAs and RAs. The 174 reports ac-
counted for approximately 11 percent of the 1,522 TCAS II is an aircraft-based airborne collision
TCAS II full-form reports in the ASRS data base. avoidance system that provides information, inde-
For reasons of maximum currency, the data set pendent of the ground-based ATC system, of the
included 55 unprocessed reports from the January- proximity of nearby aircraft. It alerts pilots,
March 1993 period. There were a total of 109 pilot visually and aurally, to potentially threatening situ-
reports, 51 controller reports and 14 reports where ations (intruders) by monitoring the position, clo-
both a pilot and controller submitted reports on the sure rate, and altitude of nearby transponder-equipped
same TCAS II incident. aircraft. TCAS II offers the pilot both TAs and
RAs. RAs are limited to vertical avoidance ma-
A coding instrument was developed to extract per- neuvers if an intruder comes within approximately
tinent information from the records. The coding 25 seconds to 30 seconds of closure.
instrument addressed several areas. The key points
are described in Table 1. Initial testing of TCAS II in 1984-85 by two major
air carriers, in a limited operational environment,
determined that “TCAS II is safe and operationally
Table 1
effective, [and] operates harmoniously in the air traffic
Key Traffic Alert and
control system.” The testing data indicated that “ap-
Collision-avoidance System (TCAS II)
proximately 50 percent of the alarms will be ‘pre-
Coding Points
ventive’ and the pilot will not deviate from his flight
P1 Pilot Use of TCAS II Advisory path. The average deviation will be approximately
P2 Pilot-initiated Avoidance Actions 300 feet (91 meters) based on a 1,500 foot-per-minute
P3 Air Traffic Control (ATC)-initiated (457 meters-per-minute) climb or descent.”1
Avoidance Actions
P4 Traffic Alert (TA)/Resolution Advisory (RA)
The reporting of TCAS II incidents increased sig-
Causal Factors
nificantly from 1990-1992 (Figure 1). The increase
P5 Pilot Communications to ATC about TA/RA
P6 ATC Reactions to TA/RA
P7 Pilot Comments About TCAS II Incident Traffic Alert and Collision-avoidance
P8 TCAS II Application: Prevented-caused System (TCAS II) Incident
P9 Evidence of Pilot/Controller Contention Reporter Breakdown
900
Source: Batelle/U.S. National Aeronautics and Space
800 Controller
Administration, Aviation Safety Reporting System s
ort 700 Pilot
p
e
Development of the coding form required a number R 600
nt
of iterations of trial codings and comparisons to de 500
validate the instrument. A team of experienced ASRS ci
pilot and air traffic controller analysts was assigned of In 400
to analyze and code the 174 reports, and the com- er 300
b
pleted coding forms were entered into a data base m 200
u
program for statistical analysis. The data was re- N
100
viewed, tabulated and summarized.
0
1990 1991 1992
Since 1956, the aviation community has attempted,
Year
in the face of widely spaced, but successive com-
Source: Batelle/U.S. National Aeronautics and Space
mercial aircraft midair collisions, to conceive and Administration, Aviation Safety Reporting System
implement an airborne collision avoidance
Figure 1
2 FLIGHT SAFETY FOUNDATION • FLIGHT SAFETY DIGEST • NOVEMBER 1993
included a total of 99 TCAS II incident re-
Pilot Use of Traffic Alert and
ports from controllers. This is the largest
Collision-avoidance System (TCAS II) Advisory
number of controller reports to ASRS on a
single topic since the August 1981 air traffic
controllers’ strike. The total breakdown of None 11 188 citations from
controller reporters on TCAS II incidents a total of 174 reports*
included 37 reports from en route control-
Evasive Turn 11
lers, 58 reports from terminal radar control-
lers and four reports from tower controllers.
Establish Visual Contact 29
The overall increase in TCAS II incident
reporting can also be attributed to the num-
ber of air carrier and business aircraft that Change Altitude 137
are installing TCAS II equipment in accor-
dance with the requirements of U.S. Federal 0 20 40 60 80 100 120 140
Number of Citations
Aviation Regulation (FAR) 121.356 and, pro-
*Note: The number of citations may be greater than the number
spectively, FAR 91.221 and 135.180. The of reports because categories are not mutually exclusive.
Source: Batelle/U.S. National Aeronautics and Space
reporting trend has continued upwards as
Administration, Aviation Safety Reporting System,
flight crews and controllers experience vari-
ous TCAS II avoidance advisory situations. Figure 2
Figure 2 depicts the actions taken by pilots in con- Reporter narratives also confirm that ATC work-
junction with a TCAS II RA based on the data set. load has increased. Pilot actions to avoid one
The majority of pilot actions were precipitated by conflict sometimes precipitated a loss of separa-
RAs (92 percent). tion or airborne conflict with another non-intruder
aircraft in the same vicinity (Figure 6, page 5).
Altitude changes ranged from 100 feet (30 meters) to
1,600 feet (488 meters), as indicated in Figure 3, There were indications in the reporter narratives
with an average altitude change of 628 feet (191 that the process of assimilating TCAS II into the
meters). In contrast, the designers of TCAS II ex- ATC system was affecting pilot-controller interac-
pected altitude deviations to average 200 feet to 300 tions. Intensive coding analysis of the 174 reports
feet (61 meters to 91 meters). Initial TCAS II simu-
lation testing did not reveal that some aircraft would
Pilot-initiated Altitude Change
deviate by 1,000 feet (305 meters) or more and in-
20 19
trude upon other occupied altitudes. Average Altitude Change
18 628 Feet
As shown in Figure 4 (page 4), the pilot informed 16
s
ATC after reacting to the RA in the majority of nt14 13 13 13
e
incidents (93). In 15 incidents, the pilot did not cid12
nthoet ipfyil oAtT eCit haebro fuotr ethwea RrnAe dm AaTnCeu (v9e)r .o rI nin 2f6o rimnceidd eAnTtCs, er of In108 8 7
while complying with the RA (17). A number of b 6
m
6
controller reports indicate that this response pattern Nu 4
4
is a major source of their concern because of the
2 2 2 2
impact of the unanticipated avoidance deviations on 2 1 1 1 1 11 1
the controllers’ air traffic situation. 0
000050000000000000
050570000000000000
112223456780123456
It is here that the dual impacts of TCAS II applica- 1,1,1,1,1,1,1,
tions can be seen. Analysis of the reporter narra- Feet
tives verify that TCAS II has been instrumental in Source: Batelle/U.S. National Aeronautics and Space
Administration, Aviation Safety Reporting
preventing many near midair collisions (NMACs)
System
and other conflicts. (Figure 5, page 4).
Figure 3
FLIGHT SAFETY FOUNDATION • FLIGHT SAFETY DIGEST • NOVEMBER 1993 3
TCAS II has definitely prevented or re-
Pilot Communications with Air Traffic Control
duced the threat of airborne conflicts as
(ATC) on Traffic Alert and Collision-avoidance
reflected by the majority of favorable
System (TCAS II) Resolution Advisory (RA)
ASRS reports and evidenced by the fol-
Ignored RA per ATC instructions 2 lowing statements from ASRS reports:
Unable to notify ATC because 4
of frequency congestion 291 citations from
Other 8 a total of 174 reports* • “It is the captain’s opinion that TCAS
Forewarned ATC before complying 9 II saved about 100 lives today. I am
Complied with RA and ignored 12 a firm believer in TCAS II forever”
ATC instructions
Pilot did not notify ATC 15 (ASRS Pilot Report No. 212377);
Warned ATC while complying 17 and,
Unknown 30
Informed ATC after RA response 93 • “Gap in updating of radar-displayed
Complied with RA — no ATC 101 Mode C misled me to think air carrier
counter-instructions
[name deleted] was still level at flight
0 20 40 60 80 100 120
level 310 (31,000 feet [9,455 meters]).
Number of Citations
Situation was resolved by TCAS”
*Note: The number of citations may be greater than the number
of reports because categories are not mutually exclusive. (ASRS Controller Report No. 225920).
Source: Batelle/U.S. National Aeronautics and Space
Administration, Aviation Safety Reporting System
However, there have been side effects re-
Figure 4
lated to the controllers’ sense of a loss of
control of the aircraft nominally under
verified that a degree of contentiousness existed in their jurisdiction. Apprehensive controllers increas-
approximately 24 percent of the reports. Figure 7 ingly wonder “Who is in control?” when TCAS II-
(page 5) shows that in a majority of the incidents equipped aircraft execute avoidance maneuvers or
(95), disagreement was not evident in the reporter’s apply non-standard uses of TCAS II without prior
narrative. However, in the 43 episodes related to notification or coordination with ATC.
TCAS II applications, and 27 episodes unrelated
to TCAS II applications, there was inferred or In a recent U.S. congressional hearing on TCAS
stated evidence of contention based on the narra- II, Air Line Pilots Association (ALPA) President J.
tive comments of the reporter. Randolph Babbitt testified that “line pilots have
strongly endorsed TCAS II and would em-
phatically resist any efforts to reduce its op-
Traffic Alert and Collision-avoidance
erational effectiveness.”2 But, at the same
System (TCAS II) Advisory Prevented
hearing, National Air Traffic Control Asso-
ciation (NATCA) President Barry Krasner omi-
Loss of Standard 69 citations from nously warned that TCAS II “is an airborne
Separation 5 a total of 66 reports* system that works improperly and erodes an
already precarious margin of safety in the
skies.”3 The pilot community, particularly ALPA,
Airborne Conflict/Less
12 sees TCAS II as a “last ditch, they-may-have-
Severe
hit-if-something-is-not-done piece of equip-
ment that gives [the pilot] a way out if the
rest of our ATC system has somehow unac-
Near midair collisions 52
countably failed.”4
0 10 20 30 40 50 60 Looking at the steps taken in implementing
Number of Citations
TCAS II, it appears that TCAS II designers
*Note: The number of citations may be greater than the number
of reports because categories are not mutually exclusive. may not have fully anticipated all the ways
Source: Batelle/U.S. National Aeronautics and Space that pilots would act, and controllers would
Administration, Aviation Safety Reporting System
react, once TCAS II was broadly implemented.
Figure 5 More than 2.5 million revenue miles later,
4 FLIGHT SAFETY FOUNDATION • FLIGHT SAFETY DIGEST • NOVEMBER 1993
with TCAS II installed in more than 70
Traffic Alert and Collision-avoidance System
percent of the air carrier fleet and with
(TCAS II) Advisory Caused
several hundred RAs reported, the follow-
ing issues have become evident:
Near Midair Collisions 2
82 citations from
Training and preparation have not been Controlled Flight 4 a total of 52 reports**
Towards Terrain
adequate. Both pilot and controller initial
TCAS II training has been inadequate for Airborne Conflict/Less 17
the unexpected avoidance actions being en- Severe
countered. A recent NTSB recommendation Loss of Standard
22
to the FAA stated that “both controllers and Separation
pilots need more training in the traffic alert
Increased ATC* 37
and collision-avoidance system.”5 There ap- Workload
pears to have been a lack of operational
0 5 10 15 20 25 30 35 40
human factors impact analysis preparatory *ATC: Air Traffic Control Number of Citations
to implementing TCAS II system-wide. **Note: The number of citations may be greater than the number
of reports because categories are not mutually exclusive.
Source: Batelle/U.S. National Aeronautics and Space Administration,
One major air carrier’s TCAS II flight Aviation Safety Reporting System
crew indoctrination video indicated that
Figure 6
resolution command altitude changes
would typically require a 200-foot to 300-foot intruders, RA commands being contradicted by
(61-meter to 91-meter) maneuver. The initial counter-instructions from controllers, etc.
FAA controller TCAS II training video tape con-
veyed that the implementation of TCAS II would A statement in the NATCA News noted that “the
be “transparent to the controllers” and the RAs GENOT (FAA General Notice RWA 3/141) is riddled
would be limited to “200 feet to 300 feet” in with ambiguities and contradictions that decrease
altitude changes. However, both FAA and ASRS the safety margin. TCAS is not reliable enough for
RA altitude-change data now confirm that the the FAA to order controllers to take such a hands-
average altitude change is more than double what off position.”8
was originally anticipated.
The FAA recently announced its intention of man-
An FAA report said “air traffic controllers have dating TCAS II software logic changes (change 6.04)
expressed major concern about the magni-
tude of the altitude displacements in re-
Evidence of Pilot/Controller Contention
sponse to some corrective RAs. The data
provide evidence that a problem exists in
Inferred by ASRS* 19
the way the pilots use the system or in the Coding Analyst 251 citations from
way pilot training is implemented.”6 Unrelated to TCAS** 27 a total of 174 reports***
Application
An ALPA statement noted: “Senior FAA CBreawse Cd oomn mFleignhtst 31
officials admitted [during the International
Based on Controller 36
TCAS Conference, Jan. 7-9, 1992, Wash- Comments
ington, D.C.] that TCAS training for con- Related to TCAS** 43
Application
trollers and controller involvement ... have
None Evident 95
been too little, too late.”7
0 10 20 30 40 50 60 70 80 90 100
Ad hoc fixes have been necessary. Subse- Number of Citations
*ASRS: Aviation Safety Reporting System
quent fixes to unanticipated TCAS II logic **TCAS: Traffic Alert and Collision-avoidance System
***Note: The number of citations may be greater than the number
deficiencies convey an after-the-fact ap- of reports because categories are not mutually exclusive.
Source: Batelle/U.S. National Aeronautics and Space
proach to the overall problems with the
Administration, Aviation Safety Reporting System
technology, i.e., nuisance RAs, phantom
Figure 7
FLIGHT SAFETY FOUNDATION • FLIGHT SAFETY DIGEST • NOVEMBER 1993 5
by Dec. 31, 1993, through the release of a notice of traffic control instructions” (ASRS Con-
proposed rule making (NPRM) and the subsequent troller Report No. 232487); and,
issuance of an airworthiness directive (AD). This
fine tuning of TCAS II software logic should • “Pilot of air taxi [name deleted] elected to
eliminate a majority of the nuisance RAs. However, respond to RA from aircraft [name deleted]
the changes will, in some instances, reduce RA warning and descended [toward] aircraft [name de-
time, and they have provoked strong reservations leted]. This is insane at best and a blatant
from ALPA and others, as illustrated below: misuse of technology” (ASRS Controller
Report No. 209177).
• “There is some industry concern that pro-
posed modifications ... will further There are indications of nonstandard use of TCAS
complicate TCAS implementation and II by pilots. The ASRS has received a number of
safety risks;”9 and, TCAS II incident reports that clearly identify non-
standard applications of TCAS II:
• Pilots are in a quandary about
how to respond to TCAS II • “I cannot have pilots using TCAS
alerts. Pilots are being placed II for visual separation. There is
T
in a dilemma about making no TCAS II separation” (ASRS
here are
split-second decisions whether Controller Report No. 202301);
to ignore controller adviso- indications of
ries on separated traffic or to • “Approach appeared to not have
nonstandard
follow RA commands. the ‘big picture.’ TCAS was used
use of TCAS II to ensure safe traffic separation”
The following statements were taken (ASRS Pilot Report No. 183286);
by pilots.
from ASRS pilot reports:
• “Cannot have pilot using TCAS
• “Center later admonished us II to maintain spacing” (ASRS
for descending from altitude Controller Report No. 202301);
and stated traffic was level … .” (ASRS and,
Pilot Report No. 205812);
• “ALPA’s ATC committee cautions pilots
• “We did not inform the controller of the ... not to use TCAS to provide their own air
evasive action in a timely manner” (ASRS traffic control and self-separation.”10
Pilot Report No. 212715); and,
The conclusions of the TCAS II incident analysis
• “ATC took offense to TCAS and its use in verified the following points:
the air traffic system. I have been finding
that ATC is not too fond of TCAS because • TCAS II has definitely enhanced flight safety
it takes away their authority” (ASRS Pilot and is widely supported by the commercial
Report No. 206966). pilot community;
Controllers feel conflict about the appropriate re- • Air traffic controllers have been confused
sponse to RA-commanded deviations. Controllers and occasionally alarmed by the variety of
have recently been instructed by the FAA to not RA applications seen in terminal and en
countermand RAs which could contradict avoidance route airspace;
commands. In some instances, the controller becomes
a passive observer to potential loss of separation • The requirements associated with the imple-
situations, contrary to all prior training and basic mentation and subsequent nurturing of TCAS
instincts as an air traffic controller: II may not have fully anticipated the “growing
pains” that have influenced both pilots and
• “TCAS should only be used as an advisory controllers;
tool for pilots and should not override air
6 FLIGHT SAFETY FOUNDATION • FLIGHT SAFETY DIGEST • NOVEMBER 1993
• Initial TCAS II training did not adequately • Direct pilots not to exceed a 500-foot (152-
prepare pilots and controllers for the sur- meter) altitude change during an RA response
prises generated during RA situations; unless the intruder is displayed to be climb-
ing or descending within 500 feet of the TCAS
• TCAS II applications have had an influ- II aircraft’s altitude;
ence on the respective roles of the pilot
and the controller; and, • Expand and vigorously promote TCAS II
simulation training for flight crews and con-
• Behavioral ramifications need to be fully evalu- trollers. The FAA training academy should
ated in concert with ergonomic issues. consider developing and distributing ATC
TCAS II situational training scenarios for
immediate use at all terminal radar and en
TCAS II Enhancements Suggested route center dynamic simulation training labs;
Information provided to the ASRS by pilots and • The FAA and major aviation organizations
controllers suggests that the following enhance- including ALPA, Aircraft Owners and Pilots
ments should be given serious consideration: Association (AOPA), American Pilots’ As-
sociation (APA), National Business Aircraft
• Mandate a requirement for high-performance Association (NBAA) and NATCA should
turbojet aircraft to reduce their vertical speed sponsor joint pilot/controller area and regional
below 1,500 feet per minute (457 meters TCAS II workshops to promote cooperative
per minute) when 1,000 feet (305 meters) dialogue between the constituencies; and,
from their assigned level-
off altitude; • ATC facility managers, through
the support of the air carriers
“F
• Require pilots to select TA and the Air Transport Associa-
or TCAS to
mode only when established tion of America (ATA), should
on the final approach for work as designed, actively promote TCAS II fa-
parallel approaches, in vi- miliarization flights for control-
immediate and
sual meteorological condi- lers. The opportunity to ride
tions (VMC) at airports correct crew jumpseat, observe TCAS II situ-
where the parallel runways ations and engage in construc-
response to TCAS
are less than 4,300 feet tive dialogue with flight crews
(1,311 meters) apart; advisories is should promote a better under-
standing of their respective pro-
essential.”
• Require pilots to notify fessional concerns.
ATC whenever a TA be-
comes a “yellow circle in- According to AC 120-55A, issued
truder” to warn ATC of a in August 1993, TCAS II is “in-
probable RA situation. A “yellow circle tended to serve as a backup to visual collision
intruder” refers to a display on TCAS avoidance … and air traffic separation service.”
instrumentation that illuminates a solid
amber circle when a TA is issued. As The AC adds: “For TCAS to work as designed,
flight crew workload and frequency con- immediate and correct crew response to TCAS
gestion permits, the pilot should also cite advisories is essential.” Flight crews, the AC said,
the clock position, direction of flight, are expected to respond to TCAS according to the
and altitude of the intruder; following guidelines:
• Require ATC to issue the pilot all Mode C • Respond to TAs by attempting to establish
traffic in his vicinity upon being informed visual contact with the intruder aircraft while
of yellow circle intruder; maintaining ATC assigned clearance; and,
FLIGHT SAFETY FOUNDATION • FLIGHT SAFETY DIGEST • NOVEMBER 1993 7
• When an RA occurs, the pilot flying should References
respond immediately “… unless the flight
crew has definitive visual” contact with 1. Haines. The Journal of Air Traffic Control
the aircraft causing the RA. The AC said (October-December 1990).
the initial vertical speed response is ex-
pected within five seconds and that alti- 2. Air Safety Week (August 10, 1992).
tude excursions “typically should be no more
than 300 [feet] to 500 feet [91 meters to 3. NATCA News (July/August 1992).
152 meters] to satisfy the conflict.”
4. Hanson. Delta MEC Safety Newsletter (Sec-
But the AC cautioned: “ATC may not know when ond Quarter 1992).
TCAS issues RAs. It is possible for ATC to un-
knowingly issue instructions that are contrary [op- 5. Aviation Daily (March 26, 1992).
posite] to the TCAS RA indications. Safe vertical
separation may be lost during TCAS coordination 6. AIRINC Research Corp. Results of the TCAS
when one aircraft maneuvers opposite the vertical Transition Program (TTP). An interim report
direction indicated by TCAS and the other aircraft prepared at the request of the U.S. Federal
maneuvers as indicated by ATC. As a result, both Aviation Administration. December 1991.
aircraft may experience excessive altitude excur-
sions in vertical-chase scenarios because of the 7. Steenblik. “How Fares TCAS? Part I.” Air Line
aircraft maneuvering in the same vertical direc- Pilot (April 1992).
tion. Accordingly, during an RA, do not maneuver
(cid:1)
contrary to the RA based on ATC instruction.” 8. NATCA News (March 1993).
[Editor’s Note: This article was adapted from Be- 9. “Interim Reports Give TCAS Mixed Reviews.”
havioral Impact of TCAS II on the National Air Airport Operations (September/October 1992).
Traffic Control System, a Battelle program report
for NASA’s Aviation Safety Reporting System, April 10. Steenblik. Air Line Pilot (April 1992).
1993.]
8 FLIGHT SAFETY FOUNDATION • FLIGHT SAFETY DIGEST • NOVEMBER 1993
U.S. Airline Industry Safety Record Improves
Despite Deregulation and Huge Financial Losses
The U.S. Airline Deregulation Act of 1978 raised concerns about
its impact on safety. Although there was an increase in fatal accidents
after its enactment, trends have improved each year since.
by
Gerard M. Bruggink
Aviation Safety Consultant
[Editor’s Note: In the January 1991 Flight Safety number of aircraft occupant deaths was less than
Digest, Gerard Bruggink first examined the safety that in scheduled Part 135 (commuter) accidents.
aspects of airline deregulation. He concluded then Financial losses, however, reached an all-time
that a series of accidents following deregulation high despite an unprecedented level of passen-
could be a harbinger of poor performance to come ger traffic.
unless the airline industry quickly came to terms
with the changes and upheaval that deregulation The airlines’ financial plight is mentioned only be-
caused. Following is Bruggink’s updated assess- cause it does not seem to fit the prevailing notion
ment based on recent accident statistics.] that poor economics tend to provoke shortcuts that
lead to accidents. The staggering financial losses did
During the three-year period 1990-1992, U.S. air- not have the corner-cutting effect typically associ-
lines operating under Part 121 of the U.S. Federal ated with economic survival because these losses
Aviation Regulations (FAR) set several records. were an inherent part of competitive strategies.
In number of departures and hours flown per The prediction that the Airline Deregulation Act of
fatal accident, this was the industry’s best three- 1978 would adversely affect air carrier safety is
year performance. And for the first time, the also being questioned. Although deregulation’s
FLIGHT SAFETY FOUNDATION • FLIGHT SAFETY DIGEST • NOVEMBER 1993 9
destabilizing influence on the work force may have public. Expressing safety in hours flown (or the
played a role in several accidents, the overall trend number of departures) per accident would satisfy
since 1978 has been one of improvement, with the that requirement. This leaves one question open:
exception of 1987-1989. That makes the 1990- Should all accidents be considered or just the fatal
1992 performance all the more remarkable and ones?
justifies an emphasis on the more intangible ele-
ments of success in accident avoidance and on the The conspicuity of fatal accidents creates headlines
individuals who deserve credit for this. that shape the public’s perception of aviation safety,
and public concern about being in a fatal accident
This article deals only with the safety performance should be approached head-on. The most visual method
of U.S. carriers because a lack of consistency in to accomplish this is to publish the number of hours
the worldwide reporting of exposure flown or departures per fatal accident.
and accident data precludes meaning-
ful comparison. According to a broad interpretation of
A International Civil Aviation Organiza-
There is no agreed-upon method to pivotal tion (ICAO) definitions, a fatal acci-
measure safety. The common prac- dent involves the death of any person
asset of the air
tice of using past accident experi- as a result of the operation of an air-
ence as a yardstick is flawed because transport craft while there was intent for flight.
it equates the absence of accidents The inconsistency in the application of
industry is the
with safety — without accounting this definition is easily demonstrated.
for the close calls and the hundreds public’s trust A refueling mishap that results in a
of daily incidents and malfunctions fire and the death of the refueler while
in the safety of
that are absorbed by the system. How- the aircraft is at the gate with passen-
ever, using those occurrences in a the system. gers or crew members aboard is treated
safety formula would not only dis- as a fatal aircraft accident. When the
courage their reporting and investi- same mishap occurs in association with
gation, but also create a bias against periodic maintenance it is not consid-
the most forthright carriers. This leaves ered a fatal aircraft accident, although
accident rates as an imperfect but credible mea- the catastrophic potential is comparable and of
sure of safety performance. concern to the traveling public.
A variety of accident rates have been used: per The same definition may skew the accident statis-
million aircraft miles; per 100 million revenue tics by not excluding work-related accidents that
passenger kilometers; per 100,000 departures; per do not threaten the safety of the aircraft. Some
100,000 hours, etc. Most of them are expressed in examples are propeller-to-person accidents involving
decimal fractions that do not convey a clear pic- ground personnel and fatal injuries associated with
ture of year-to-year changes. A more informative activities at the gate, such as the towing or push-
yardstick would be the use of hours flown per back of aircraft. These industrial mishaps should
accident, which is nothing but a variation of the not become part of the fatal aircraft accident sta-
rate per 100,000 hours. To convert such a rate into tistics just because they happened while there was
the number of hours flown per accident, divide intent for flight. Once they go into the statistical
100,000 by the rate. For example, an accident rate blender, these accidents assume the same weight
of .05 per 100,000 is the equivalent of 100,000 = 2 as the collision of two Boeing 747s in Tenerife,
million hours per accident. The same method is Canary Islands, and distort the true risks of flying.
used to calculate the number of departures per
accident. The U.S. air carrier accident experience presented
here is derived from annual statistics compiled by
A pivotal asset of the air transport industry is the the U.S. National Transportation Safety Board
public’s trust in the safety of the system. This (NTSB). Exposure data (hours flown and the num-
means that the industry is best served by a perfor- ber of departures) were provided to the NTSB by
mance yardstick that is readily understood by the the U.S. Federal Aviation Administration (FAA).
10 FLIGHT SAFETY FOUNDATION • FLIGHT SAFETY DIGEST • NOVEMBER 1993