A Comparison of FAA Integrated Noise Model Flight Profiles with Profiles Observed at Seattle-Tacoma Airport [Exerpts]

by
George W. Flathers, 11
December   1981
MTR-81W288
SPONSOR: Office of Environment and Energy
CONTRACT NO.:DTFA01-82-C10003
PROJECT NO.:1494A
DEPT.:W-47
This Document was prepared for authorized distribution.
It has not been approved for public release.
The MITRE Corporation
Metrek Division
1820 Dolley Madison Boulevard
McLean, Virginia   22102
[Editor's Note: This Document is available to the U.S. public
through the National Technical Information Service, Springfield,
Virginia 22161]

[Page v]
EXECUTIVE SUMMARY
Federal Aviation Administration's integrated Noise Model (INM) Is
a series of computer programs designed to assess the noise impact
of aircraft operations in the vicinity of an airport.  The user of the
INM supplies data concerning the airport and runway layout, the
number and types of aircraft, and description of the flight tracks
they use. the IN computes and reveals the noise environment In
terms of preselected noise metrics of the user's choice.  As part
of MITRE's overall effort to check the validity of the results of
INM computations for the FAA, a comparison was made between
the arrival And departure profiles contained in the INM data base
and those observed in actual operations at the Seattle-Tacoma
International Airport.  A flight profile describes aircraft altitude and
velocity as a function of distance from the runway during a takeoff
or an approach to landing- The extensive data base In which the
INMprofiles are stored also contains noise and other performance
data for various types of aircraft.
In the spring and autumn of 1979, MITRE conducted a similar
flight profile study which to presented In MTR-80WO0119,
"Comparison of FAA INM Flight Profiles with Observed Altitudes
and Velocities at Dulles Airport,"  Reference 1. The main
conclusion of that study was that, for departure operations, most
airlines were using procedures which differed significantly from
those assumed by the then-current Number 7 INM data base.  Due
to the sparsity of data sampling locations and limitations in the data
collection mechanisms, however, the exact nature of observed
departure profiles could not be determined.  Since the time of
the Dulles study, the FAA has prepared a new data base (Number
8) which includes revisions based an a relatively recent FAA
Advisory Circular (AC91-53, Reference 2) outlining recommended
standard noise abatement departure procedures.   The comparisons
in the present study are made with respect to the Number 8 INM
data base.
Methodology
The basic approach taken in this study is an extension and a
refinement of that taken In Reference 1. ARTS-III radar data
provided the raw information upon which statistical inference, could
be made about actual flight operations.  Using the target reports
provided by the ARTS-III system, and a special smoothing
technique called cubic spline function smoothing, the altitude and
velocity of each aircraft on arrival or departure was determined over
several

[Page vi]
points within 10 nautical miles of the airport- The altitudes and
velocities of  nearly 3000 arriving or departing aircraft were
determined in this manner from data collected in the period between
May and July, 1981. This large sample was aggregated into smaller
samples according to the type of operation conducted (namely,
arrival or departure) and the type of aircraft involved.  In this study,
sample sizes were large enough to permit investigation of the
following six aircraft types: DC-9. B-737, B-727, DC-10, L-1011,
and B-747.  Profiles for each type of aircraft were characterized
statistically and compared directly with appropriate profiles taken
from the Number 8 INM data base.
Results for Arrivals
The INM approach profile for standard air carrier arrivals depicts a
continuous vertical decent along a 3 degree glide slope to the point of
touchdown approximately 1000 feet beyond the runway threshold.
The speed of the aircraft within 10 nautical miles of the runway
is assumed to be constant at the INM supplied final approach speed.
When compared to this profile, the following trends were noted:
 	Observed altitude profiles suggested that all six types of
aircraft closely follow the 3 degree glide slope.  The usual sources of
descent guidance for an air carrier pilot on an approach to landings
are the Instrument Landing System (ILS) glide slope. or an optical
aid called the Visual Approach Slope Indicator (VASI), both of
which provide an approximately 3 degree glide slope.  Observed altitudes
varied around the glide slope as a function of distance from the
runway: as Aircraft approached the runway, variations In
observed altitude became progressively smaller and more centrally
distributed about the 3 degree glide slops.
 	Observed velocity profiles revealed that most aircraft were
performing a decelerating approach rather than one of constant
speed.  Most aircraft approaching the airport are at a significantly
higher speed but slowed to within a few knots of the INM
designated final approach speed as they came within 2 nautical
miles of the runway.  The frequent occurrence of the decelerating
Approach Is consistent with the predominant conditions at Seattle:
VFR weather and fairly light traffic, both of which make
decelerating approaches practical.

[page vii]
Results for Departures
There are  many other factors associated with departures which
contribute to considerably more variation in observed operations.
There are procedural differences in the way the departures are
performed by various airlines.  In addition. there are performance-
limiting factors such an aircraft weight, pressure altitude,
temperature, and wind which introduce additional sources of
variation.   Accordingly, a more detailed analysis of departures was
performed.
All airlines specify their own standard departure procedures for
their flight operations manuals.  These procedures are usually
fashioned After the FAA suggested noise abatement departure
profile. (as outlined in FAA Advisory Circular 91-53, Reference 2),
with various levels of compliance.  The profiles of most airlines
resemble each other for aircraft with high bypass ratio engines.  For
low bypass ratio engines, however, the FAA procedure specifies a
greater thrust reduction after takeoff than some airlines use.  This
would result in a steeper climb angle than under the FAA
procedure, with all other factors hold equal.
The INM data base, on the other hand, has a set of completely
defined profiles for each aircraft type which were constricted under
the assumption that the FAA procedure is being followed by all
aircraft. In addition, this data base has up to seven slightly
different profiles for each aircraft type to reflect differences in
departure performance attributable to varying departure weights.
Under the assumption that aircraft departure weight and stage-
length (the non-stop fight distance) are proportional. the INN
estimates departure weight by using stage length as an Index.  The
profile for the most likely stage-length was used as the INN
baseline for the comparisons and the following results were noted:
	Observed altitude profiles for the DC-9 sad B-737 were
much lower than the INM profiles for the near field segment (the
portion of the departure within 3 nautical miles of the Brake
Release Point (BRP)).  There was fairly close agreement between
observed and INM profiles for the other aircraft in the near field
segment.  For the far field segment (the portion further than 3 n.m.
from BRP) the DC-9 and B-727 were much higher than the INM
profiles.  A possible reason for this observation is that the
procedures used by the pilots of these two aircraft types are not
fashioned after the FAA profile which the INM assumes.
 [Page viii]
	Observed velocity profiles were within reasonable
agreement with INM profiles for the near field segment for all six
types of aircraft. For the far field segment all observed velocity
profiles were close to the INM profiles, with the exception of the 3-
727 which was faster than the INM profile.
1)	An analysis of observed B-727 departures was performed to
determine if differences in departure procedures of different airlines
have observable effects on Actual departure performance.  The
median B-727 departures of five major airlines were compared with
each other.  No real differences were observed In altitude profiles
of the five airlines for the departure segment within 5 n.m. from
BRP.  Beyond this point. however, the disparity became more
distinct.  At 8.5 n.m. from BRP the highest median departure was
approximately 1000 feet higher than the lowest.  There were no
tangible differences in the velocity profiles for the entire departure.
A review of available flight operations material revealed that the
airline with the lowest median attitude at 8.5 n.m. also employs a
sharp thrust cut-back which was ultimately intended by FAA AC91-
53.  The expected and observed result of this cut-back wits the
shallower climb angle.
 	To measure the sensitivity of both observed and INM
profiles to differences in stage length, on analysis was performed on
B-727 departures grouped into four different stage-lengths.  It was
found that slight but palpable differences exist in both INM and
observed profiles due to stage-lengths.  However, variation from
other sources to several times greater than the sensitivity of the INM
to changes in stage-length.
 	Based on the findings of this study, the FAA Office of
Energy and Environment proposed a few revisions to the Version 8
INM departure profiles for the DC-9, B-737, and B-727.  The
revised profiles were the result of recomputing departure
performance based on the departure procedures which were
evidently in use by pilots of these aircraft. The agreement of
observed profiles with the revised Version 8 profiles was
found to be significantly improved.
Page ix]
Conclusions and Recommendations
This profile study represents the most comprehensive comparison
made to date between observed operations and profiles contained In
the INM data base.  In general, the new Number 8 profiles have
significantly improved observed-INM profile agreement.  Because
the version of the INM which implements the Number 8 data base
had not yet been released, the sensitivity of noise estimates to
differences Iii flight profiles was not investigated.  This sensitivity
should be quantified in n future effort.  However, it is anticipated
that improvements in the flight profiles will, in most instances,
result in more accurate noise estimates.  Major observations, and
recommendations to make the INM easier to use and to improve the
accuracy of results.are listed below:
 	For arrivals, the agreement between observed operations and
likely INN profiles was generally good.  Observed arrivals for ail
six types of aircraft followed the 3 degree glide slope and exhibits
decelerating approaches.
 	At present, the Number 8 data base contains predefined
approach profiles which describe approaches of constant speed for
the last 10 nautical miles before touchdown.  The predictable
patterns of observed arrivals at Seattle-Tacoma suggest that
inclusion of a decelerating profile in the data base may also be of
benefit to the user, especially at locations where weather and traffic
conditions make decelerating approaches popular.
 	For departures, observed INM profile agreement was good
for aircraft with high bypass ratio engines. but the agreement was
not so good for low bypass ratio engines.  The disparity for the case
of low bypass ratio engines was attributed to differences between
assumptions under which the INM profiles were constructed and
actual operating practices used by various airlines.  This hypothesis
was supported by the analysis of  B-727 departures grouped
according to airline which indicated that differences In observed
profiles could be traced to proceduraldifferences.  The revised INM
profiles for the DC-9., B-737, and B-727 proposed by the FAA
result In significantly improved agreement with observed profiles,
and they should be incorporated as a permanent part of the INM
data base.
[Page x]
The analysis of B-727 departures grouped according to stage-
length revealed that differences between INM profiles for the
shortest And longest stage-length tend to be marked by variation
from other sources. In addition, the assumption that weight
estimation can be based on stage-length categories may not always
be true.  Based an these findings the number of stage-length
categories should be reduced from a maximum of seven to a
maximum of two or three.
[Page 5-1]
5.  CONCLUSIONS AND RECOMMENDATIONS
This analysis of aircraft profiles represents the most comprehensive
comparison made to date between observed operations and profiles
contained in the INM data base.  A more Complete review of
airline operating practices has been included to reveal those
operational variables which are likely to Influence the shape of
observed profiles.  In general. the new Number 8 INM data base
profiles have made significant Improvements in observed-INK
profile agreement.  There are, however, a few areas where the
agreement could be improved even further and INN ease-of-use and
efficiency enhanced.
For arrivals, the agreement between observed operations and
standard INM approach profiles vas generally good.  The standard
INM altitude profile depicts a continuous descent on a 3 degree glide
slope to the point of touchdown.  Observed arrivals for all six types
of aircraft were closely grouped around this glide slope.
A difference was noted, however, in the comparison of observed
and INM velocity profiles for arrivals.  Standard INM velocity
profiles depict an approach of constant speed for the last 9 nautical
miles before the threshold. Observed aircraft, however, approached
the airport area at a significantly higher speed and gradually
reduced speed to the final approach speed approximately 2 nautical
miles from the runway threshold.  Thin observation was attributed
to the prevalence of weather and traffic conditions which made
decelerating approaches feasible.
Though the INN user could construct his own decelerating
approach to accommodate such a situation, the predictable patterns
of observed operations suggest that the addition of a completely
predefined decelerating approach would be more efficient.
consistent, and of greater benefit to the user.  The user would
have to be informed of the weather and traffic conditions which
make either the constant speed or decelerating approach applicable,
but the benefit gained in establishing this choice is the simplicity in
which the user can specify entire approach profiles which are based
on predictable and fairly invariant observed operations.
Another issue concerning decelerating approaches to the effect they
have on estimated and observed noise levels.  The thrust values
contained in the INM approach profile data base are specified for
aircraft maintaining a given configuration in a
[Page 5-2]
"steady state".  An aircraft which is decelerating, however, is not in
a steady state and is probably using less thrust than an aircraft
maintaining a constant speed in the same configuration.  The end
result of the decelerating approach should be some reduction in
noise generated at the source.  However, the size of the noise
reduction may be small because thrust levels are generally low even
in constant speed approaches.
For departures, the comparisons made between INM profiles and
observed operations showed little difference for some aircraft types
and greater differences for others.  In general, observed INM profile
agreement was better for wide-bodied aircraft with high
bypass ratio engines.  The close agreement was attributed to the
similarity between assumptions under which the INM profiles were
constructed and actual operating practices used by various airlines.
The observed INM agreement was not quite as good for narrow-
bodied low bypass ratio engined aircraft.  For the near-field
segment of the departures, INM profiles for the B-737 and the DC-
9 were much higher than observed operations.  The INM profiles for
these two aircraft for this segment reflect rather steep climbs which
are probably not attainable in everyday operations.  On the far field,
differences were noted for the DC-9 and B-727.  The observed
trends suggest that the thrust cutbacks in actual operation are
not as great as those assumed by the INM profiles.  Some airlines
employ a departure which specifies a smaller thrust reduction than
the FAA departure for low bypass ratio engines.
The analyses of B-727 departures grouped in various ways also
contributed to a greater understanding of the pertinent variables
involved in departures.  An analysis of B-727
departures grouped according to airline revealed that some
differences in observed profiles could be traced to differences in
operating procedures.  Another analysis performed on B-727
departures grouped according to stage-length resulted in the same
conclusions an when they were considered in aggregate.
Differences between INM profiles for the shortest and longest
stage-lengths are not great and tend to be masked over by variation
from other sources.  In addition, the assumption that weight
estimation can he based on stage-length my not be true for all
instances.  Based on these findings the number of stage-length
categories should be reduced from seven to two or three.
The revisions to the INM profiles proposed by the FAA for the DC-
9.  B-737, and B-727 aircraft resulted in such improved observed-
INM profile agreement.  The revised profiles were the
[Page 5-3]
result of recomputing departure performance under different
assumptions about the departure procedures being used.  To
guarantee that the INM profiles maintain relevance with general
observed operations, the revised profiles should become a
permanent part of the INM data base.
In conclusion, the INM profiles contained In the new Number 8 data
base generally agree with current observed profiles.  The level of
agreement to much better than afforded by the older Number 7 data
base.  However, the improvements suggested above would lead to
even closer agreement and ease the tasks presented 
to the INM user.
[ Editor's Note:  A search of government documents December
1995 found no supplemental studies or reports verifying the results
of the FAA INM noise modeling program have been validated
through calibration of the noise modeling program with radar data at Seattle-Tacoma International Airport since this study.]