High resolution airport airspace model using geographic information system

Material Information

High resolution airport airspace model using geographic information system
Panayotov, Apostol Panayotov
Publication Date:
Physical Description:
iv, 145 leaves : ; 28 cm


Subjects / Keywords:
Airports -- Geographic information systems -- United States ( lcsh )
Aeronautics -- Geographic information systems -- United States ( lcsh )
Runways (Aeronautics) ( lcsh )
Air traffic control -- Simulation methods ( lcsh )
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )


Includes bibliographical references (leaves 142-145).
General Note:
Department of Civil Engineering
Statement of Responsibility:
by Apostol Panayotov Panayotov.

Record Information

Source Institution:
|University of Colorado Denver
Holding Location:
|Auraria Library
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
527639140 ( OCLC )
LD1193.E53 2009d P36 ( lcc )

Full Text
Apostol Panayotov Panayotov
Master of Science, Moscow Railway Engineering Institute, 1992
Master of Engineering GIS, University of Colorado Denver, 2006
A thesis submitted to the
University of Colorado Denver
in partial fulfillment
of the requirements for the degree of
Doctor of Philosophy
Department of Civil Engineering

(2009) by Apostol Panayotov Panayotov
All rights reserved.

This thesis for the Doctor of Philosophy
degree by
Apostol Panayotov Panayotov
has been approved
Iliya Georgiev
Jacek Grodecky
Wesley Marshal

Panayotov, Apostol P. Doctor of Philosophy, CU Denver Civil Engineering PhD
Thesis directed by Professor Lynn E. Johnson, PhD
This thesis develops innovative methods for high-resolution airspace modeling
needed to conduct airport runway airspace analyses. The Federal Aviation
Administration (FAA) Federal Aviation Regulation Part 77 (PART 77)
establishes standards and notification requirements for objects affecting navigable
airspace. The overall methodology presented in this paper is called the GIS-based
Airspace Analysis Model (GAAM). The GAAM allows for rapid and accurate
computation of the specific 3D airport airspace governed by FAA PART 77
regulations. The method for PART 77 model development is termed GIS database
from survey data (GISFSD) and is based on high accuracy spatial data.
Presented processes and tools are based exclusively on existing ArcGIS
capabilities, which avoids extensive programming and third party software
implementation. To suit better 3D representation and because ArcGIS does not
support true 3D volumetric modeling this paper propose a new approach of
development and generating a pseudo 3D model for airspace analysis
(P3DMAA). Airport planners can now accomplish these analyses in a timely,
technically accurate and FAA regulatory responsive manner.
This abstract accurately represents the content of the candidates thesis. I
recommend its publication.

I dedicate this thesis to my wife, Tzvetanka and my son Plamen for their
unfaltering support and understanding while I was completing this thesis.

My thanks to my advisor, Professor Lynn E. Johnson, for his contribution, advices
and support to my research. I also wish to thank all the members of my committee
for their valuable participation and insights.

TABULAR FORM...............................................20
FIGURE 3: OE/AAA QUAD MAP......................................21
FIGURE 5: TEXT REPORT OF RESULTS...............................23
FIGURE 6: DRAWING OF IMAGINARY SURFACES........................24
FIGURE 8: PRINTABLE REPORT OF 2D IMAGES........................25
FIGURE 10: CAD LAYERS..........................................46
FIGURE 12 ACCURACY EXAMPLE.....................................58
FIGURE 13: SURVEY DATAFILE.....................................60
FIGURE 14: WAYS OF MODEL DEVELOPMENT...........................61
FIGURE 15: RESEARCH METHODOLOGIES..............................64
FIGURE 16: PROPOSED RUNWAY DATABASE............................65
FIGURE 17: CAD TO GIS RESULTS..................................66
FIGURE 18: CAD TO GIS FLOWCHART................................69
PART77 MODEL...............................................71
FIGURE 21: PARAMETERS..........................................73
FIGURE 22: PARAMETERS APPROVED BY FAA..........................74
AND EXTEND.................................................79
FIGURE 26: PS SURFACES.........................................80
FIGURE 27: SIGPOINTS AND SURFACES..............................82
FIGURE 34 OTHER CRITICAL SURFACES..............................94
FIGURE 35: 3D DATABASE STRUCTURE...............................97

TABLE 1: DISTANCE AND DEVIATION ANGLE.......................53

1. INTRODUCTION.........................................................1
1.1 General overview....................................................1
1.2 Issues surrounding current methods for airport analyses.............4
1.2.1 PART77 problems and facts.........................................6
1.2.2 Research hypothesis...............................................7
1.3 Aims of research effort.............................................8
1.4 Contribution of Thesis..............................................8
1.5 Scope and limitations...............................................9
1.6 Thesis overview.....................................................9
2. REVIEW OF AIRPORT ANALYSES..........................................11
2.1 Airport Traffic Control Tower Line of Sight Analysis...............11
2.2 Terminal Instrument Procedures.....................................12
2.3 Height Limitation Study............................................12
2.4 Objects Affecting Navigable Airspace...............................12
3. LITERATURE REVIEW...................................................16
3.1 General overview...................................................16
3.2 Airspace analysis software.........................................17
3.2.1 Currently available airspace analysis software.................17
3.2.2 Planning Technology, Inc. Three-Dimensional Airspace Analysis
Program 17
3.2.3 CGH Technologies -- Obstruction Evaluation.....................19
3.2.4 Federal Airways & Airspace Obstacle Evaluation and Management
Systems 21
3.2.5 ClearFlite Airfield Obstruction Identification Tool..........26
3.2.6 PLTS for ArcGIS An Aeronautical Solution.....................27

3.2.7 Airspace Modeling by Crawford, Murphy and Tilley, Inc..........28
3.2.8 Summary...........................................................28
3.3 3D Modeling.........................................................29
3.3.1 Overview of 3D modeling...........................................30
Octree-Coded Model..............................................30
Constructive Solid Geometry Model...............................32
Wire Frame Mesh Model or TIN....................................33
TIN/TEN Models..................................................33
3.3.2 Modeling schemas..................................................34
3.4 Geographic Information System.......................................35
3.4.1 GIS concepts......................................................36
Triangulated Irregular Network..................................37
Terrain Model...................................................38
3.4.2 Spatial databases.................................................39
3.4.3 GIS for 3D........................................................40
3.5 High resolution measurements and data capture methods...............41
3.5.1 Imagery...........................................................41
3.5.2 LIDAR and HDS.....................................................42
3.5.3 GPS...............................................................43
3.5.4 Traditional ground survey methods.................................43
4. PROPOSED AIRSPACE ANALYSIS MODEL.....................................45
4.1 Overview............................................................45
4.2 PART77 and ARAA as general processes in CAD.........................46
4.3 FAA specifications..................................................49
4.4 3D Mathematical foundations.........................................49
4.5 GIS methodology and 3D spatial representation.......................54
5. DATA COLLECTION / ACQUISITION:.......................................55
5.1 Existing Ground surface model.......................................55
5.2 Existing manmade structures model...................................56

5.3 PART77 Model
6. GIS BASED AIRSPACE ANALYSIS MODEL.............................61
6.1 Overview.....................................................61
6.1.1 Variant 1 Full PART77 model................................62
6.1.2 Variant 2 Combined individual PART77 surface models to form a Full
PART77 model.....................................................62
6.2 PART77 model development from existing CAD drawings..........64
6.3 Development from scratch using initial surveying data (GISFSD).72
6.4 Full airport PART77 model....................................87
6.5 Verification and sensitivity of the model......................88
6.6 PART77 and other protective imaginary surfaces.................94
6.7 The 3D model.................................................95
7. CONCLUSIONS AND RECOMMENDATIONS...............................99
7.1 Implementation...............................................99
7.2 Contribution.................................................99
7.3 Contribution to the professional community..................101
7.4 Problems encountered........................................102
7.5 Results.....................................................103
APPENDIX A -CAD to GIS modeling....................................106
APPENDIX B Primary Surface....................................113
APPENDIX C SIGPOINTS and surfaces.............................117
APPENDIX D Airport Survey.....................................135
APPENDIX E Surface Parameters.................................138

APPENDIX F Model builder flowchart to create one runway PART77 from

1. Introduction
1.1 General overview
In the last several years, the Federal Aviation Agency (FAA) has
published a series of very important documents that herald important changes in
aviation transportation and pose a significant impact to the airport transportation
business. These documents will force airports to move forward in the
technological realm, mandating the collection of highly-accurate spatial data,
developing better and more accurate airport layout plans (ALP), master plans
(MP), and 3-dimensional (3D) airport models, and employ state-of-the-art
geographic information systems for running their business.
The FAA documents to which I am referring are the following advisory
AC 150/5300-16A (09/15/2007) General Guidance and Specifications
for Aeronautical Surveys: Establishment of Geodetic Control and
Submission to the National Geodetic Survey. This document explains the
specifications for establishing geodetic control on or near an airport. It
also describes how to submit information to the National Geodetic Survey
(NGS) for approval and inclusion in the National Spatial Reference
System (NSRS) in support of aeronautical information surveys.
AC 150/5300-17B (09/26/2008) General Guidance and Specifications
for Aeronautical Survey Airport Imagery Acquisition and Submission to
the National Geodetic Survey
AC 150/5300-18B (05/21/2009) General Guidance and Specification
for Submission of Aeronautical Surveys to NGS: Field Data Collections
and Geographic Information System (GIS) Standards
A Guide to Airport Surveys (05/15/2009). This document encompasses
16A, 17B, 18B, and Airport GIS (AGIS) and provide guidance to
complete data collection process.
These documents describe the means of how to capture and store spatial
data and develop an airport GIS database, particularly an electronic Airport

Layout Plan (eALP). They further indicate that the FAA has established a
centralized data repository to store the airports eALP and together with the
aforementioned documents will serve as a foundation for the development of
enterprise airport geographic information system (AGIS).
Additionally, the FAA is currently working on its NextGen
implementation plan, targeting the development of avionic equipment with the
goal of moving airspace traffic management from being ground-based to satellite-
based. This new way of traffic management will rely solely on satellite controls
and, consequently, will require the development and support of highly-accurate
3D airport models that not only ensure airport safety but also provide cost-saving
Further, these changes will, without a doubt, impact airspace and
navigational aids (NAVAIDS) analyses as well. While it is true that airport
planning and development sections perform some of airport airspace analyses
related to runway safety, the FAA conducts the most important analyses and will
continue to do in the foreseeable future.
With the implementation of new and advanced technological methods and
devices, the quality of data capturing has gained precision and will continue to do
so. Global Positioning Systems (GPS) technology, LIDAR, satellite imagery, and
high definition surveying have made the process of data capturing easy-to-execute
with highly accurate results. New software applications have also been developed
and implemented that not only more accurately and efficiently process data and
analyze the results, but now also automate the preparation of sophisticated
presentations for non-technical audiences. With these technologies currently in
place, we can say that the stage is set for the FAA to require more airspace and
NAVAIDS analyses to be made by airport planning sections. And most
importantly, by giving more freedom to airport planning and development
divisions, the federal agency will in the long term save money while fulfilling the
role of expediting the airport development and design processes, not to mention
that this hands-off approach will provide airports with the latitude to significantly
improve their airport operations.
As airports moves toward the implementation of NextGen and AGIS, the
Airspace and NAVAID obstruction analyses applications remain high-priorities
for the airport planning and development sections as these analyses are based on
different spatial data models. Depending on the type of analysis conducted, often
two or more spatial models are simultaneously involved. The most important of
these are the models of existing manmade structures (EMSM), existing ground
surface (EGSM), NAVAIDS data, and the PART77 model. In this case, I am
referring specifically to spatial models, all of which can be loosely defined as 3D
models, but are not in a technical sense actual three-dimensional models. In fact,
currently, most airports store two-dimensional files, and airport runway airspace

analyses are still conducted using an older method that does not support true 3D
spatial models. This fact is based on Airport Survey Summary presented in
Appendix D Airport Survey. This survey was conducted with intention to
determine how other airports in US store their Airport Layout Plan (ALP), Mater
Plan (MP) and run necessary airspace analyses.
However, contrary to the current trend, there are a few airports in USA
participating in an FAA pilot program. It is the goal of this program to develop
and test the use of spatial databases and eALP and develop a means to greatly
integrate GIS into the everyday operations of an airport. Currently, there is little
use of GIS across airports in the U.S., and most of the existing GIS applications
that are in use have little relationship to true airspace analyses. This is probably
because not many airports consider GIS an appropriate tool for running
complicated airport analyses. Most aviation professionals, and especially these in
management positions, consider GIS to be a robust mapping tool or property
management program, partly because the tool is advertised and promoted more as
a business, mapping or management tool. However, in reality, professional GIS
users know that GIS is an extremely powerful software application and its
analytical capabilities are some of its strongest features.
Essentially, if we examine the three previously mentioned models
involved in airport airspace analysis, we can see a direct cause and effect
relationship between the data used to create these models and the actual airspace
and NAVAIDS analyses themselves. Without a doubt, data accuracy is one of the
primary requirements for buildings spatial models. However, the presence of
highly accurate data does not ensure that the chosen model is, in fact, the correct
model. Consequently, the use of the wrong model for airspace and NAVAIDS
analyses could result in costly analysis operations that yield less than useful
results, gaining very little. Therefore, the second most important requirement in
spatial analysis should be model correctness, and several important factors come
into play when choosing the right model. These must include at a minimum: data
accuracy, strictly defined model development methods, and the possibility for
reiteration of the modeling process to test for the same results. To ensure that
spatial models are effective in meeting the goal of providing high quality results
in airspace analyses, it is imperative that they also provide: 1) a well-defined
process for model development, so the model can be executed efficiently and
effectively and reiterated as needed, and 2) the process not require a specialized
team of consultants to execute and maintain.

1.2 Issues surrounding current methods for airport analyses
Let us take for example a typical scenario in which construction activities
are proposed in or around an airport. In such a case, the airport typically first
submits form 7460 -1: Notice for proposed construction and alteration to the
FAA Airport District Office (ADO) to notify the FAA of the proposed
construction. Then, an airspace analyses are conducted to determine the
construction activities or heavy equipment use, such as cranes, that are in close
proximity to the runways and that could potentially violate the FAA standards and
recommendations regarding penetration of any of the airports protective
imaginary surfaces. Following this review, the FAA issues a letter of decision
with an Approved, Objected, or Conditional determination status plus any
recommendations for addressing potential issues or violations.
As the 7460 process can take from two to four months to complete, it is
often too costly and time-consuming for airports to undertake. To mitigate the
cost and expedite the FAA review process, typically airport planners perform
extensive in-house analysis to predetermine any potential violations to FAA
regulations that may arise in regard to the proposed development and
construction. In so doing, they are able to construct multiple alternative solutions
and submit them together at one time when filing form 7460-1, resulting in a
significant saving of time and money.
The most frequently used model by the airport planners for this type of in-
house analyses is the PART77 model. This model allows for the representation of
all protective imaginary surfaces described by FAA in Title 14 Code of Federal
Regulations (CFR) PART77 Objects Affecting Navigable Airspace.
Traditionally, the PART77 map is created in a computer-aided design (CAD)
environment, employing certain CAD techniques and taking in account only
runway end points, while simultaneously disregarding the points that describe the
runway centerlines. While some airports use 3DAAP and OMS software, most
airports prefer to have their own PART77 map as simple CAD drawing. The
process of developing a complete imaginary surface map for a single runway,
using computer-aided design (CAD) normally requires numerous hours of labor
and a multitude of calculations, with the majority of calculations computed
outside of the CAD environment, requiring the resulting computed data to be
manually re-entered into CAD. This approach is not only prone to human-error,
but if an airport has multiple runways (most commercial airports typically average
three runways), the difficulty of manually calculating data is only further
compounded. After the imaginary surface maps for each runway are created, the
next step is to combine all the imaginary surfaces into a one that depicts the whole
airport. This secondary operation, in turn, requires many hours of computer

calculations, also computed outside the system, and, thus, also requiring manual
reprocessing within the system. Thus, the process may require days to complete a
full PART77 imaginary surfaces map for an airport. Subsequently, if there are any
changes to original data, the process would need to be entirely repeated, thus,
doubling an already arduous task. Later, when the map is finished, it is used for
airport runway airspace analyses (ARAA).
This example methodology presents some significant problems. It
supports a process that is cumbersome, inefficient, and prone to human error.
Moreover, because of the extensive time required to execute the entire process,
maintaining and updating the PART77 map as changes occur in the airport can be
difficult to impossible. Additionally, there are other significant issues with airport
runway airspace analyses just described.
First, typically the information gathered in the analysis has a target
audience of business and government stakeholders, that can include municipal
and county officials, airline representatives, construction companies, other
affiliated vendors, and the general public. This audience seeks data that supports
business and environmental goals and objectives and lacks the technical
orientation required of the airport analyses process that ensures a clear
understanding of the results. In other words, this audience requires that an
additional step be added to the process, a translation of the technical data and
concepts into a presentation format that is more transparent and readily digestible
by this non-technical audience.
Second, to meet the FAA requirements for PART77 development, a series
of documents must be used, PART77 [1],[2],[3],[4],[5]. However, no direction or
recommendations currently exist with the FAA to assist airports in the completion
of PART77. While this lack of documentation provides airport planners with the
freedom to select their own means, tools, and methods for developing and
analysis model and conducting the analyses, often this fosters an ad-hoc
development environment that breeds inconsistencies and make replication of the
process virtually impossible. Due to the lack of specific guidelines, the chances
are that two different planners will create a PART77 model using different
techniques and different interpretations of how to build the model, resulting in
maps or models that are quite diverse. And, coupled with this is the potential for
inconsistent choices made for the data used to develop PART77, that is, only
runway centerline points, only runway end points, or something entirely different.
The result is different input data will obviously output distinctly different models.
Additionally, not every type of airspace analysis requires the use of all
imaginary surfaces that are part of PART77. This further raises the question is
it necessary to develop a full PART77 model or develop each imaginary
protective surface as a separate surface model? This distinction is an important
one and must be made at the onset of the development project and not put off to

when the results are available. It is extremely useful to have both a complete
model of the entire airport as well as a separate dataset for each runway, enabling
planners to manipulate multiple combinations between surfaces during the
analysis process. The downside of this approach is that distinct models require
more data space and an extended development time, not to mention better naming
convention for each imaginary surface in the project.
Finally, it has become apparent that in the not too distant future, existing
two-dimensional (2D) CAD PART77 maps will not be able to integrate
effectively with eALP and AGIS and will require an entirely new generation of
3D airspace planning tools. Moreover, we need to contain costs and execution
time when maintaining, updating, or replicating PART77 and other necessary
The solution lies within the new technologies and software that are
addressing better and more reliable ways to store, retrieve and manipulate spatial
data. The next sections will present a summary list of discussed problems and
facts and in so doing lay, a foundation for the research hypothesis presented in
this paper.
1.2.1 PART77 problems and facts
1. Existing PART77 map are based on runway end points only
2. Existing PART77 map are not sufficiently accurate to meet the newest
FAA requirements
3. Updating existing PART77 maps is inefficient and often
4. Present PART77 maps do not comply with the latest FAA ACs and do not
support the process of developing eALP and AGIS
5. Supplemental calculations are manually computed outside the main
software application used for PART77 development
6. Data is required to be re-entered manually
7. PART77 map is usually 2D or 2.5D CAD drawing
8. CAD drawings and CAD software are not suitable for conducting
sophisticated spatial analyses
9. Airspace analyses are routinely conducted at airports, as they are an
essential means for planning airport development and construction
10. There is currently no FAA guidance or standard that describes unique or
general procedures to do some specific airspace analyses performed on
daily basis by airport planners

11. Every Airport has its own unique way of developing airspace analysis
models, often using in-house or specific proprietary software or a mixed
combination of software packages for analyses and PART77 development
12. Often due undefined procedures, it takes from several hours to several
days to complete one or more airspace analyses
13. The presentation of the final results are overly technical and difficult for
non-technical business and community parties to interpret and understand
the community impacts
14. Presently, the drawings, maps, and products do not support What-if
15. Often, current methods require development team to implement more than
one software application to complete a single analysis, requiring the
purchase of multiple systems and vast skills sets for team members to
manage the job tasks
16. There are many analyses that are customer driven. These analyses need
more flexible data models.
1.2.2 Research hypothesis
In this paper, 1 will propose an entirely new airspace analysis model that is based
on GIS instead of drawings and will increase the accuracy and speed of the
airspace analysis process. I further propose a new database design and PART77
model design to be completed in GIS without using additional third-party
software, and that all analyses be run within a GIS environment, especially within
ArcGIS, without requiring extensive software developing. I expect that the new
model will require some customizations but not to the extent of developing a
completely new software package. Rather, the customizations will take the form
of a new tool that will be implemented as a module within ArcGIS itself and to
allow for the creation and representation of 3D models in GIS environment.
Lastly, I propose that the presentation of final analysis results will be made in a
format that is more easily understood by a non-technical audience, and that the
new proposed type of modeling be designed for use with not only PART77 but
also for other restrictive imaginary surface models.

1.3 Aims of research effort
This research will focus on the investigation and development of a defined
process for the development of a geodatabase and airport spatial PART77 model,
specifically. The exclusive focus on this model is because the airport planning and
development personnel routinely and almost daily use PART77 to conduct some
of the most important airspace analyses. It is important to note that these airspace
runway analyses are conducted from the perspective of an airport planner and not
a pilots perspective. These types of analyses are specifically conducted to help
airport planners and developers design new airport structures, schedule
maintenance of existing infrastructure, or planning for future airport expansion.
Other models, such as those that address existing ground and manmade structures,
will also be discussed, as they are an integral part of airspace runway analyses.
However, it is important to understand that the development of these latter models
are not the central focus of this research project.
To address these issues, this paper proposes the use of a GIS database and
developing spatial PART77 model. The goal of this research is not to develop a
new 3D modeling theory or new software based on that theory. Clearly, with
substantial resources, any enterprise or individual can develop new theories and
software. Rather, the intent of this paper is to develop a practical and
straightforward solution to automate the development process of complete airport
spatial PART77 model and to streamline the process of airspace analysis based on
existing ArcGIS capabilities and its extensions.
1.4 Contribution of Thesis
The contribution to the field of GIS will be the method itself and the benefits it
provides for technical users of the ArcGIS package and airport planners. The
proposed method for this PART77 model development is original and unique.
First of all no one has done this before. The approach I am using is unique with its
exclusive use of GIS environment without integrating or using third party
software. The presented method is streamlined and automated and is virtually free
of user mistakes. The significance is that this method can be used for further
development of other imaginary restrictive surfaces models, it provides high
accurate model which increase the accuracy of airport runway analyses that are
essential for airport safety operation and right management decisions.

My objectives in meeting this goal will be:
1. To ensure ease-of-implementation by enabling the process to be stored as
a Model Builder procedure and/or a Python/VB script;
2. To run the model development process within ArcGIS, using its existing
tools and extensions;
3. To present a database model flexible enough to support user updates and
customization as necessitated by specific airports requirements;
4. To contain the cost of implementation by the use and reuse of currently
implemented technologies rather than providing an entirely new design or
integrating third-party software solutions;
5. To eliminate maintenance fees or yearly-based contract for data and model
6. To minimize human interaction during the development process and thus
eliminate personal mistakes
7. To use a single file data entry for full PART77model development
8. To assure that the repeatability of the process will have the same result
1.5 Scope and limitations
The scope of work includes development of geospatial database for data
storage and development the method for automation and streamlining the process
of PART77 modeling.
PART77 is based on 3D spatial data, stored and developed as TIN/Raster
model. Due to present limitations of existing ArcGIS software PART77 cannot be
presented as full volumetric 3D. However, the presented PART77 model will
allow for running three of the most important ARAA in 3D. It is not a target of
this research to develop full airport volumetric 3D model.
1.6 Thesis overview
The research project schema of this dissertation consists of five steps describing
milestones of the research process and presents the final PART77 model.
Step 1 Planning: Investigate possible ways to develop a database that will hold
the data necessary to create PART77 model.

Step 2 Data Collection/Acquisition: Research and compare existing data types,
methods and techniques of data collection.
Step 3 Methods and methodology: Develop mathematical equations and logical
procedures used for PART77 model development and use real data examples to
test the process.
Step 4 Results: Verification of the sensitivity of the database structure, models
and process
Step 5 Conclusions: Draw the final conclusions and outline future improvement
and recommendations for proposed method implementation
Methods, problems and solutions presented in this thesis are real and presently
used in Denver International Airport for developing PART77 and TERPS spatial
models as well as conducting ARAA on a daily basis.

2. Review of airport analyses
Airport runway analyses are routinely conducted at airports, as they are an
essential means for planning airport development and construction. Although
there are a number of different types of analyses, all of them share a single goal:
keeping navigable airspace clear of obstructions. Their purpose is to ensure that
all planned airport development and construction comply with the regulations and
standards put forth by the Federal Aviation Administration (FAA), and in so
doing, guarantee that air traffic is free of any penetration or obstacle in navigable
Specifically, the airport runway airspace analyses must observe the FAA
standards and requirements and policies and procedures outlined in Title 14 Code
of Federal Regulations (CFR) PART77, United States Standards for Terminal
Instrument Procedures (TERPS), Advisory Circular (AC) 150/5300-13- Airport
Design, and AC 150/5190-4- A model zoning ordinance to limit the height of
objects around airport and pay particular attention to Airport Traffic Control
Tower Line of sight (ACTC-LOS) and height limitation in airspace.
The following section discusses the types of airport analyses most
frequently conducted by airport planners. These include Airport Traffic Control
Tower Line of Sight Analysis, Height Limitation Study, and some Terminal
Instrument Procedures. For purposes of this paper, we will term them airport
runway airspace analyses (ARAA) so that we can distinguish ARAA from the
airport airspace analyses conducted by the FAA.
2.1 Airport Traffic Control Tower Line of Sight Analysis
The airport traffic control tower line of sight analysis (ATCT-LOS)
ensures that any proposed construction or alteration to any airport surface does
not affect an air traffic controllers visibility from the air traffic control tower to
any part of the runway-taxiway complex.

2.2 Terminal Instrument Procedures
Another type of analysis that is used less often by airport planners is the
Terminal Instrument Procedures (TERPS). The basis for these procedures is
founded on the criteria used by the FAA to develop Approach and Departure
procedures for civilian and military airports in United States (US). TERPS
ensures that no single object penetrates the approach and departure surfaces.
However, because of the infrequent use of TERPS by airport planners, only two
of the important TERPS surfaces will be incorporated into the proposed model.
TERPS, in general, will not be targeted in the model design.
2.3 Height Limitation Study
The Height Limitation Study (HLS) ensures that a newly proposed
construction object stays well below the limitations established by PART77 and
the Terminal Instrument Procedures (TERPS) and, at the same time, meets the
requirements of the ATCT-LOS analysis. This analysis includes the location of
heavy equipment used for construction.
For example, if the use of a crane is identified as part of the specifications
for completing a proposed construction, then the analysis that determines the
optimal location for placing the crane would include an estimate of the maximum
allowable height of the proposed structure. Further, if the heavy equipment has
the potential to penetrate one or more of the imaginary surfaces, the FAA
regulations would require proper marking of the equipment and/or runway
shutdown during construction. Thus, this analysis must take into account the
location of any heavy equipment to be used during construction to ensure that
they do not incur problems for airport operations during the construction process.
2.4 Objects Affecting Navigable Airspace
All of the aforementioned analyses are based mainly on the FAA
regulation (CFR14 PART77 Objects affecting navigable airspace). This
regulation defines the imaginary surfaces around the airport and sets standards to

control the height of objects located near and around the airport using a map that
delineates the restricted areas around the airport. There are six surface types for
each runway: Primary, Horizontal, Conical, Transitional, and Approach. The
approach surface can be further defined as having two parts: 40:1 and 50: 1.
If there is more than one runway in an airport and the runways are each of
a different type, then two imaginary surface maps are developed for each runway,
and the two maps are combined for use in the final analysis. This process is time-
consuming and quite difficult, as the first step for developing the imaginary
surface map for each runway requires an analysis of the six types of imaginary
surfaces just described. Then, the two maps are combined and the intersection of
the surfaces is calculated, resulting in the final map. This process meets the
regulation outlined in the FAA regulation PART77, requiring that a combined
imaginary surface map be created for each airport wishing to undergo
construction. To create this map, consulting companies typically use CAD and its
accompanying software features, as it is a very difficult to render the map
The imaginary surface models from airport to airport and from runway to
runway differ markedly. This is due to the fact that each surface type must meet
the specifications put forth by the FAA PART77. An example of the specification
is shown in Figure 1.

federalAviaicn F^sals!ki5Psrt77!A)ve>idix B
t- iftcrrv HtflMi
- iiMurr Win MUnanM ** IM[
MirtT iU
rntf tfHMi kmi PM
MUM ttfeUtttrt aMIHU
/_.At ****-<>
' ttcTim >-
1 imi civil. mi imiiut iimcli
Sou ice: Ffc&st Alston Fbgi/stitis P3ft7?
FAR Part 77 Imaginary Surfaces

Thus, running an analysis for each separate runway and for each particular
surface is not only a common practice among airport planners, it is often a
necessity. Further, to run a complete analysis of the airspace, airport planners
must not only include the models of imaginary surfaces as described by FAA in
documents [1] to [6] from the references, but also include the terrain surface,
buildings, antennas, radars and any other man-made structures that exist above
ground level.

3. Literature review
3.1 General overview
Before we can begin the discussion of current software features that would
serve the needs of airport planners when performing airspace analysis, it is
imperative first to define the models that we are planning to create, and define the
set of needs that will work as the criteria that will evaluate the effectiveness of the
current software.
An Existing Surface Model (ESM) is created from aerial images, satellite
images, contour maps, and survey data, or any combination of these to enhance
the accuracy and details of the model. This model consists of two sub-models:
existing manmade structures model (EMSM), and existing ground surface model
A PART77 or other Protective Imaginary surface model (PISM) is created
from series of rules and dimensions as specified in the relevant FAA
Additionally, the ability to import data files and manually enter data is
required to build the models. The model-development method must support
automatic and semi-automatic model types: the automatic model-development
follows an established logical model and is based on mathematical equations, and
the semi-automatic model-development is operator-driven and follows well-
known and established logical procedures.
Other requirements include:
Airport ownership of the data models themselves in order to efficiently
manage updating the models
Support for data changes and updates in 2D models, as well as 3D
Support for multiple types of analysis, such as ATCT LOS, HLS, and
What-If scenarios
Support for multiple types of presentation, such as a text-based format, 2D
and 3D color printouts with data
Ease-of-use, excluding the need for special equipment or specialized user
skills, a single environment, GIS, to include all process operations from

data entry to report generation, and the use of a custom application for all
process operations as well
Ease-of-implementation, excluding the need for extensive programming,
the development of large add-in modules, or the use of additional third-
party software
Manageable and reasonable implementation costs, excluding costly on-
going maintenance contracts and software upgrades
3.2 Airspace analysis software
3.2.1 Currently available airspace analysis software
To assist airport planners in completing the necessary airspace analyses,
there are several off-the-shelf software applications currently on the market. Each
once can perform some of the airspace analyses that this paper describes. The
following paragraphs discuss the most frequently used programs and the
advantages and limitations of each type.
My goal in this discussion is not to critique the existing software
applications that have been professionally developed, but rather to present how
my work can aid in the extension and expansion of the current capabilities of
existing analysis software.
3.2.2 Planning Technology, Inc. Three-Dimensional Airspace
Analysis Program
The Three-Dimensional Airspace Analysis Program (3DAAP) was jointly
developed several years ago by the Florida Department of Transportation and
Federal Aviation Administration and Planning Technology, Inc. This application
assists planners in more effectively determining whether proposed construction
projects can cause potential interference with the surrounding airspace. One of the
distinct advantages of using this program is that it performs a combined analysis
on the proposed development as required by PART77, TERPS, LOS, and
obstruction shadows [29],

3DAAP was designed to operate inside a CAD environment, using
AutoCAD Map versions 2005 and 2006 and Micro Station as the base platform. It
does not build a true 3D surface model. Analysis calculations are based on
trigonometric rules not on model assessment. This method ultimately limits the
capabilities for analysis and presentation. As CAD based application it is not
capable for simultaneous multi feature analysis. Based on the airport survey
conducted earlier and presented in Appendix D Airport Survey, the results
shows that the airports currently using 3DAAP typically contract annually with
Planning Technology, Inc. to maintain a Web-based portal where analysis is
conducted through the use of their proprietary software. Because the airport does
not own or maintain the base data or airspace models that drive analysis when
using 3DAAP, any change that affects the imaginary surface model must be
reconfigured in the application by Planning Technology, Inc. prior to beginning
any type of analysis. This forces the customers of 3DAAP to maintain a
relationship that requires on-going support of the base software and client data
and only serves to further extend the timeframe for running an analysis.
Although the software is freeware, an initial implementation can be time-
consuming and costly. For example, at Denver International Airport (DIA),
installing and configuring the base 3DAAP software required several iterations to
get their 2006 version of AutoCAD to operate with 3DAAP. After the
implementation was complete, it ran for one week, and then failed. There was no
apparent reason for the failure, and the implementation was abandoned.
Further, 3DAAP for airspace analysis is unable to represent analysis
results using a 3D model. It simply provides printouts and reports of quantitative
data in the form of text files, severely limiting the ease and efficiency of data
interpretation for a non-technical audience.
Although 3DAAP is one of the most-used programs in airspace analysis,
the facts remain that the solution is proprietary, and the software company
ultimately owns the analysis model and the underlying code to the model,
resulting in a time consuming and inflexible solution for airport planners. Further,
this prevents airport planners from extending the model beyond its present
capabilities and tailoring it to meet the needs of future analyses.

3.2.3 CGH Technologies Obstruction Evaluation
Obstruction Evaluation is a process that determines whether a proposed
object or alteration penetrates imaginary surfaces, and if this object poses a
potential risk as an obstacle to navigable airspace.
To meet this need, a Web-based application was developed by CGH
Technologies for the FAA to update and automate the Obstruction Evaluation
(OE) process for objects with a height of 200 or more feet above the ground. This
application provides collection and conversion data to support the OE process. It
houses OE data through a geographic information system (GIS), AML, and
Avenue to automate the data conversion process, and it provides standard
mapping capabilities and evaluates OE cases.
CGH Technologies claims that through Geographic Information Systems
(GIS) and a Web-based intranet application, CGH is creating a more efficient and
reliable process for obtaining data, plotting geographic coordinates, performing
measurements, and ultimately, making decisions [20]. In actuality, the
application has some serious limitations.
First, it evaluates all 2D penetrating objects in 3D [19], and then reports
each penetration using a text-based format. All calculations occur behind the
scenes, running off the FAA server. This setup results in the following
disadvantages: 1) because the imaginary surface maps are not accessible and fully
dependent on the information the FAA server has recorded for the airport in
question, the resulting analysis may be skewed or inaccurate because of dated
FAA airport data, 2) the web-based Obstacle Evaluation application does not
automatically incorporate the data from previously conducted analyses, especially
the ATCT LOS analyses, and 3) the evaluation is calculated using the principles
of trigonometry and does not employ a sufficiently robust 3D model of the
imaginary surfaces.
Figures 2 and 3 illustrate the data output of the OE application model. In
essence, the software plots the evaluated facilities on Quad maps and outputs the
data in tabular format.


The presentation is not very accurate as all data is entered manually and
the analysis is run based on user-defined information that can cause user errors.
However, the resulting presentation is insufficient to meet the needs of airport
planners, as it does not render an image of the actual terrain surface. Further, it is
inadequate as a tool for planners as it does not enable planners to address the
What-If scenarios that are so often a part of planning for airspace development.
Finally, the use of Quad maps restrict a planners ability to achieve timely and
accurate analyses, as Quad maps are typically not updated frequently enough to
reflect the most accurate picture of a current airport environment.
3.2.4 Federal Airways & Airspace Obstacle Evaluation and
Management Systems
There are three Obstacle Evaluation and Management Systems (OMS)
packages currently available from Federal Airways & Airspace: Airspace OMS,

Airspace Government and Airspace Manager. Each application is customized to
meet the needs of the customer, and each package includes Airspace, Airspace
Survey, and TERPS Professional [21].
In these applications, key obstruction information, based on FAA Forms
7460 and 7480, are added and stored for analysis. Then on-screen queries of the
airspace surfaces and the obstruction points provide a precise analysis of the
impact of the obstruction on the airport airspace [21].
The Airspace Obstacle Analysis software is used to analyze and assess
potential impacts to an airport's airspace from aerial obstructions, using a dialog
box that requires manual entry of the data, as shown in Figure 4. Figure 5 presents
the results as text report.
AIRSPACf & 2004
File Chock Analysis Results Help
*3 ^ T
D G?
Airspace Analysis
5= wmi&azmm !>

Fie Name:
Ground Elevation: 1
726 AMSL CM)
Structure Height:
Overall Height:
166 AGL |H)
092 AMSL JM)
Cky (Lockbourne
| OH
Statute Mie*
Degrees (true)
D egreet: Degrees: ^
Minutox: l4e Minutes:
Seconds: |59.6 Seconds: |25.3
| 39.8165555555556 | 82.9570277777778
DNE FAR 77.13(a)(1)- EXCEEDS FAR 77.13(a)(2).

Nearest pubfic fadlaty: ILCK: RICKENBACKER
Distance to the ARP: 18287 It | 1.3638 ran.
Direction to ARP: (97.45 degrees
Distance to Runway: |BD47 K- | .9952 *>-


AIRSPACE 2004 [Analysis: \\Msp\airspace\User2YTUT0RIAL.APT] 0(D
| @) File Edit Font Airport Antenna - 1 & x |
[si & *3 T BTff= f
ID & M it % H o
| Open Public Airport Report
LATITUDE: 39"-481-59.6"
LONGITUDE: 82 -57 '-25.3 *'
Tills facility has at least one runway over 3,200 feet in length.
Your structure DNE FAR 77.13(a)(1) but EXCEEDS FAR 77.13(a)(2) Notice
Criteria for this airport. You must notify the Federal Aviation
Administration using a FAA Form 7460-1 a minimum of 30 days prior to your
construction start date. As a minimum, please review reports for FAR Part
Obstruction Surfaces, Air Navigation and Communication facilities.
EXCEEDS FAR 77.13(a)(2) Notice Criteria by: 96 feet.
You are 6047 feet from the nearest runway threshold and the threshold
elevation is 736 feet. Please review runway analysis for remaining
airport surfaces.
This airport has both Circling and Straight-In Instrument Procedures.
Dlaaga votm ot.t i T a vm iwal f twririiafVi PvnPQfiiivQc f nr
This application performs the required calculations and generates reports
that clearly identify any impact the proposed, constructed, or alteration project
may have on navigable airspace [22],
The Airspace Survey software is developed on a GIS platform and links
data analysis functionality to the geographic mapping, enabling a geo-coded
image to be exported to any number of alternative locations to Airspace OMS for
analyses. Images are equivalent to a 2C standard [23], which does not provide
sufficient accuracy, as shown in Figure 6.

-jf .m
TERPS Professional is used to analyze airport terminal approach
procedures and the height restrictions they impose [24], This application
automatically calculates and draws imaginary surfaces and performs various
TERPS calculations as ILS, LOC, VOR, etc., as shown in Figure 7.

In all cases, the applications claim to perform the majority of the
necessary airspace analysis and generate the needed report types and datasheets
resulting from the analyses. Additionally, the software also allow planners to
attach USGS Quad Maps or download aerial and satellite images from Terra
Sever at no cost and attach the images as background. Customers can order
single-or multiple-point analyses, using their website on a case-by-case basis with
analysis results delivered online or by conventional mail. FAA Form 7460-1 can
also be automatically generated through the application, a great improvement over
other comparable software packages. Printable reports, as the one shown in Figure
8, provide quality 2D images that are also available in graphical format.
It is true that the OMS systems provide a number of the much-needed
tools for efficient and effective analysis, as with other proprietary solutions
discussed earlier in this paper; however, changes to the specifications of airport
operations require reconfiguration by the software provider in order to update the
analysis models and base data. Again, this dependency limits airport planners
from running efficient and timely analyses.
Further, despite a host of usable features, there remain a number of
deficiencies in the web-based OMS:

A model that includes existing surface and buildings to ensure a full
analysis of the data is not available
All location and elevation data require manual entry
A Terrain model is not currently available
Aerial and satellite images and USGS Quad maps are static and can be
only used as background
A 3D model is not supported
3.2.5 ClearFlite Airfield Obstruction Identification Tool
ClearFlite, developed by BAE Systems, is a digital mapping tool
developed to help users in aviation industry view 3D stereo images of runways
and airfields, automatically generate Obstacle Identification Surfaces (OIS) for
single and multiple runways, identify and measure airfield and runway
obstructions, and export data to third-Party GIS and 3D visualization applications.
This tool comes with the SOCET SET photogrammetric package. ClearFlite is
used by photogrammetric and engineering organizations as an extension to their
normal mapping activities. This tool allows them to collect the obstruction data
The software uses digital photogrammetry techniques to create a digital
terrain model from stereo images, LIDAR, survey data, GIS databases, airborne
and satellite images. An Obstruction Identification Surface (OIS) is created and
based on the runway end points. Both terrain and OIS models are available and
viewable three-dimensionally. The software has the capability to compare DTM
and OIS and automatically highlight all obstruction that penetrates the OIS
The ClearFlite operator visually identifies terrain and features that
penetrate through the OIS. Using the standard ClearFlite collection tools, the
operator digitizes and records the X, Y, and Z coordinates of the entities, together
with the basic attributes. For a typical airport runway, the identification, collection
and attribution of the obstructions can be accomplished in approximately 8 to 10
hours. [25]
ClearFlite is ideal for data collection, identification of penetrations, and
creation of DTM and OIS models, all of which facilitate the creation of 3D
models. It is a very powerful program; however, it falls short of the mark when
providing analytical capabilities, and, further, requires extensive work time to
collect the most important information in airspace analysis-penetrations.
BAE Systems stated that:

ClearFlite is also used by entities that traditionally have had no
photogrammetric expertise, but have the desire or requirement for quality control
and management of the obstruction surface information[25]
It is not right, as ClearFlite requires knowledge and skill in digital
photogrammetry to work with specific tools, such as the 3D photogrammetric
Finally, to receive optimal benefit when working with ClearFlite,
professionals must purchase additional tools besides the software. An investment
in specific photogrammetric modules, special equipment such as special graphic
cards, monitors and eyewear are also warranted to enable a stereoscopic view that
is essential to operators working with stereo images.
In conclusion, based on above-mentioned facts this software package
seems primarily designed for consultants or private companies and is
inappropriate for airport planners.
3.2.6 PLTS for ArcGIS An Aeronautical Solution
The ArcGIS Product Line Tool Set (ArcGIS-PLTS) is an aeronautical
solution developed by the Environmental Systems Research Institute (ESRI). This
application includes tools and functionality to manage aeronautical information
and produce database driven aeronautical charts [26] and is an extension of the
ArcGIS Desktop.
The framework of PLTS consists of four integrated modules: Job
Tracking, Foundation Tools, GIS Data Reviewer, Map production System Atlas.
The PLTS Aeronautical Solution supports NGA, FAA and ICAO chart standards.
It is based on the principal of automated cartography to create aeronautical charts
in plans and profiles for terminal procedures in different formats, sizes, and
scales, using a broad variety of datasets.
Additionally, it provides a robust data creation tool called Future Builder
that enables users to create mapping features that are not part of the off-the-shelf
database. The mapping program provides a set of analysis capabilities: defining
the highest point in air-traffic zones using minimum vector attitude, 3D NAVAID
reception/coverage, and aircraft tracking in real-time or based on historical
position. Again, PLTS is another very powerful and useful tool. However, it was
designed primarily for NGA, enabling them to rapidly design, create, and print
aeronautical charts, while observing TERPS. It expedites the creation of accurate

and massive airport GIS databases. This package has tremendous benefits for
airports, when used in that context.
3.2.7 Airspace Modeling by Crawford, Murphy and Tilley, Inc.
An airspace modeling application in CAD for TERPS and PART77 was
developed by Crawford, Murphy and Tilley, Inc. (CMT). This software imports
airport and runway-specific airspace models directly into GIS for analysis. They
also provide a custom-built analysis tool within GIS that can look at multiple
surface analyses for a single point or cluster of points when performing TERPS
and PART77 analysis. It was originally developed as a compatible land-use model
and was not specifically designed for airport planning needs. The CMT
application worked with GIS Solutions, Inc. to write code inside GIS to allow for
the merging of runway airspace model data into multiple, airport-wide runway
CMT is currently developing a TERPS modeling and analysis software for
basic, straight-in approaches for airport clients. Future projections indicate that
they may also incorporate TERPS and PART77 surfaces into a comprehensive
analysis suite. Finally, CMT has indicated that it may package their product in
stages: one that enables model building, and another that provides the GUI-based
analysis tool that is already available in their Airspace Analyst Software.
The latter package, when it becomes available, appears to be most
appropriate for airport planners. However, its release has not yet been announced,
so an evaluation of its practical effectiveness as an airport planning analysis tool
remains to be seen.
3.2.8 Summary
A quick assessment of the software selection parameters outlined in this
section clearly indicates that none of currently available software applications
satisfies all of the essential criteria established in 3.1 General Overview to act as
an efficient and effective tool for ARAA.
The aforementioned software applications derive the analysis results,
using simple trigonometric equations based on the runway end-point data. The
coordinates of the evaluated point is usually presented in degrees, which

significantly decrease the accuracy as 1 arc second equals 30m This method
works if the accuracy is set to 1A, meaning a tolerance of 20feet/6meters
horizontally and 3feet/lmeter vertically (as based on Accuracy Standards for
Aeronautical Studies).
The fact is that no one of the software package presented here enables
creation or manipulation of a 3D airport runway airspace model. The absence of
3D data features severely constrains the analysis and the presentation of results.
The methods used in existing software applications are inadequate because they
employ not state of the art methodology, utilize seemingly inaccurate data, and do
not meet the latest standards and requirements outlined in the most recent
advisory circulars [32], [33], [34], [35], [36].
By evaluating the limitations of current software packages available on the
market, I do not mean to imply that they have absolutely no relevance to the task
of performing airspace analyses. To the contrary, many of the aforementioned
programs are currently in use today. However, I would like to stress that they lack
sufficient intelligence, flexibility, and the level of accuracy needed to be used
exclusively when performing airport analyses for airport planning. To do so,
existing software applications would need to be extensively reworked to
adequately comply with eALP and AGIS, and to meet the growing needs of
airport planners to provide for fast and accurate analysis and presentation of
results to non-technical audiences in an easy-to-understand way.
But, developments of this type are lengthy and expensive, so it may be
some time before such an application is produced. In the meantime, airport
planners have a job to do, and this research presents an alternative and new way
of developing an airport airspace model for airport runway airspace analyses
based on geographic information system (GIS) that will help them do their job
more efficiently and effectively.
3.3 3D Modeling
Anyone performing a search in Google on 3D model or 3D software
will quickly discover that the search yields more than ten million hits. However,
as a researcher probes deeper into the data, he or she will quickly discover that
most 3D modeling software is built for a specific need, providing information on
techniques that are suitable for game modeling, CAD/CAM environments, or
models directed at architectural and engineering problems. There are, in fact,
limited reports for 3D modeling or models created exclusively for the GIS

environment, and limited information on existing 3D cities, airports, or
topographic models.
Although this field is at present relatively undeveloped, it is gaining
momentum as a growing number of scientists and technologists concentrate their
efforts on developing more sophisticated and complex 3D models. We currently
see this growth trend in the extensive software development for existing GIS
platforms and the implementation of different models and modeling schemas.
3.3.1 Overview of 3D modeling
Three-dimensional modeling has had a very long history in human
development, and we see this represented most clearly in two areas of human
study: science and art. Since the beginning of human history, we see evidence of
ancient peoples trying to recreate the images of their surrounding three-
dimensional environment as accurately as possible. Initially, simple artifacts such
as stone and chisel were used, later paper and ink or pencil took their place, and
today we harness the power of technology to recreate 3D images using computers
and software.
Many different approaches to 3D modeling exist; however, ArcGIS does
not currently support all of them. The short list include the octree-coded model,
the constructive solid geometry (CSG) model, the wire frame model, the
Triangulated Irregular Network, and Tetrahedronized Irregular Network surface
models. Each of these models has a different approach and utilizes use of different
data types.
Octree-Coded Model
The octree-coded model is commonly used for spatial indexing and
collision detection. It is not an actual 3D-model, but rather a mathematical
description that presents a hierarchical model in which the smallest parts are 3D
cubes or voxels. This representation is a data structure with internal nodes that can
have up to eight children each. Boolean operations can be performed to run
logical calculations or queries. Although GIS does not currently support this
model, the model can be implemented as Python, .Net or VB script, and an Octree
method with spatial indexing can dramatically improve GIS spatial queries.
The use of the octree model enables analysts to immediately see how
surfaces intersect with one another. Consequently, this model serves as one of the
best methods for collision detection. The downside of this model is that the data
results on the collision reports are very rough and ultimately require more

extensive and sophisticated programming to ensure that the data granularity is
sufficient for accurate airspace analysis.
In one of my earlier research projects related to airspace analysis, I
considered this model as a potential way of running airspace runway analyses
because it provides a way for collision detection. The goal was that every 3D
object needed to be fully defined in the modelling cube, and each object defined
as a family of k+1 pairs of ordered pairs {P, Ej}, where 0 < i< k (k is the
maximum level), P is a finite set of properties, and Et is the set of the octants (EO
is the whole modelling cube). The property P could be geometrical, physical, or
any other that characterized the modelling object. The geometrical property of
every octant can be Empty (E), Partly Occupied (PO), or fully Occupied (O). The
octants in Et are nodes of an octree with the following features:
the root of the octree is level 0 and presents the whole modelling cube;
all other nodes can have 8 children that are ordered according to their
number in a sequence from 0 to 7;
the edges of the of the tree represent a parent-child relationship and
every octant can only have one edge to its parent;
all nodes with the geometrical characteristic Empty or fully Occupied are
leaves of the tree;
the nodes Partly Occupied from the internal levels have 8 children.
Figure 9 illustrates the octree to level 2 for a simple 3D object. If the 3D
modelling needs a more precise spatial approximation of the object, more levels
could be added [42],
e po ,'e; po (e) ;e) : e"; (e; level 1
}i :ji 6 (§) o E; E -
LEVEL 2 E O E E E E i t

This idea proved to be very good, but required a long time to compute and serious
hardware power to achieve high accurate results.
Constructive Solid Geometry Model
With its origins in CAD, the constructive solid geometry (CSG) model
deals with developing and creating solids. This model allows the creation of
complex surfaces or objects, using Boolean operators, such as union, intersection
and difference. It also uses the basic 3D primitives with varying parameters:
spheres, cubes, and cylinders [10]. It is by far the best model for the
representation of solids, outweighing all other models, such as the wire frame
mesh or octree model.
The special properties of CSG solids allow mathematical operations that
are not possible with an arbitrary polygon mesh. CSG solids are made out
of intersecting coplanar faces. Each face has an ascertainable plane
equation, defined as a 3D vector, plus a distance from the origin. [38]
Because of this, basic, intermediate and advanced operations can be conducted.
Distance between point A and plane
B = (Bd-Bnx *Ax-Bny *Ay-Bnz *Az) [38]
Intersection of ray AB with plane C:
u (- Cnx *Ax-Cny *Ay-Cnz *Az+d)/(-Cnx *(Ax-Bx)-Cny *(Ay-By)-Cnz *(Az-Bz))
x = Ax+(Bx-Ax) *u
y = Ay+(By-Ay) *u
z = Az+(Bz-Az)*u [38]
Intermediate: Solid solid intersection or Extrude solid from point or vector
Advanced: Texture mapping [38]
Texture coordinates tell the GPU how a texture lies across the surface of
an object. Solid texture coordinates can be calculated on the fly, with user control
over parameters like scale, position, rotation, and shearing. This is done by
generating texture mapping axes per face, based on face normals. The texture
coordinates for any vertex of a face can be calculated from the vertex position
andface texture mapping axes:

u = unx*x+uny*y+unz*z+ud
v vnx*x+vny*y+vnz*z+vd [38]
The positive aspect of this model is that it can be rendered using textures for
better surface sides representations. The downside is that is difficult to implement
in small computer units. Further, 3D GIS may use CSG, but database
management of the CSG primitives is lacking and little research and development
currently exists in this area. The result is that GIS cannot support CSG modeling
without the addition of a script or implementation of a new algorithm [9].
Presently, ArcGIS supports a multipatch feature, which can be used to reconstruct
a CSG model in Arc Scene for visualization. However, using this type of feature
for calculation is quite painful and does not provide required flexibility.
Wire Frame Mesh Model or TIN
The wire frame mesh is a method that illustrates 3D objects as a series of
lines that represent the edges of an object. The image is created by specifying
each edge of the physical object where two mathematically continuous surfaces
meet or by connecting constituent vertices using straight lines. The representation
of a wire frame mesh is a TIN (Triangulated Irregular Network) surface model
based on Delaunay triangulation method. TIN is a 2.5D model as it actually
represents the surface of an object and does not have any volumetric
representation. The method can be extended to include three-dimensional
modeling by using an extrude function or specifying the base elevation. TIN
surface supports volumetric calculations as cross-sectional areas and volumes, if
two or more TINs are present. This method is most suitable when modeling
terrain heights, as the Earths surface has a single height value at every X, Y
location. The main drawback of the 2.5D model is its incapacity for modeling
vertical faces and multiple surfaces at the same X, Y location [14].
TIN/TEN Models
In the last decade, the Tetrahedronized Irregular Network (TEN) model
has grown in popularity. A number of scientists from Delft University of
Technology in Delft, Netherlands have published numerous research papers and
theories implementing the TEN model in 3D urban and city modeling projects (S.
Zlatanova, Oosterom P., Penninga F., Kazar B.). Peter Van Oosterom and Friso
Penninga explored further theories presented earlier by Pigot S., Carson E. about
three-dimensional topological modeling. Edward Verbree describes a method for
better surface representation based on TEN and a 3D analyst algorithm. In
essence, the algorithm was applied to the TEN model to retrieve the TIN model
with the goal of representing the surface as accurately as possible [8].

This raises the question What is the TEN model and why it is so
important as a model data structure for 3D GIS? TEN is composed of nodes,
edges and faces and tetrahedrons. The relationships between the simplexes are
well defined in this true 3D volumetric model. The simplex is a four-sided
(tetrahedron) 3D primitive in which every three nodes lay on the same surface.
TEN is a complex model, and it is best suited for developing real 3D city or
topographic models. The distinct advantage of TEN is that it does not ignore the Z
values of each node. Thus, it is possible that it could take many tetrahedrons to
construct one factual object [10]. This is a distinct challenge for the next
generation of 3D-GIS development. However, if developed, this feature would
enable GIS to handle numerous, simultaneous analyses and ultimately offer true
3D functionality in lieu of the 3D visualizations it offers today.
3.3.2 Modeling schemas
The last several years has produced a quantity of research in the
development of automated methods for 3D object reconstruction. Although the
flurry of development has introduced a variety of different approaches, I have
singled out four methods as being directly applicable to the discussion put forth in
this paper regarding 3D airport modeling. They are the bottom-up method, the
top-down method, the detailed reconstruction method, and the combination
method as presented by Stoter [10] in his 3D GIS, where are we standing?
Bottom-up: using footprints (for existing 2D maps) and extrudes, the
footprints with a given height use laser scan data, or survey, GPS, or
photogrammetric data. The problem with this approach is that the details
of roofs cannot be modeled. Since one value is usedfor every footprint, the
buildings appear as blocks in the model. The approach, however, is very
fast and sufficient for applications that do not need high accuracy (do not
need roofs or the roofs are flat and leveled) and many details.
Top-down: using the roof obtained from aerial photographs, airborne
laser scan data and some height information from the ground (one or
more height points near buildings, DTMs). These approaches emphasize
the modeling of the roofs [11][12]. Obviously, the accuracy of the
obtained 3D models is dependent on the resolution of the source data.
Detailed reconstruction: The most common approach is to fit predefined
shapes (building primitives) to the 3D point clouds obtained from laser
scan data [13] or 3D edges extracted from aerial photographs [12] [13].
The advantage of this approach is the full automation and the major

disadvantage is that it is very time-consuming since the algorithms used
are very complex.
Combination of all of them e.g. lasers scan data and topographic data
(Hofmann et al 2002), aerial photographs and maps [13], etc.
An analysis of the current methods and practices leaves much to be
desired. Clearly, no single schema stands out above the others as the most
effective and productive method for modeling three-dimensional structures for
airspace analysis. While one method is quick to execute, it results in insufficient
accuracy. Another method provides greater detail, but is costly and time-
consuming to generate. Yet another provides insufficient accuracy and does not
meet the requirements outlined for this project.
For the purpose of accurate airport runway airspace analyses, we need to
have full 3D models of the airport plus 3D models of PART77 imaginary
surfaces. Clearly, developing the needed topographic models and 3D model of
man-made objects and creating imaginary surface models is not only time-
consuming, but also presents a very challenging and difficult process. Developing
PART77 map and adding some of TERPS surfaces in 2D are easily done in a
CAD environment as we can apply the principles of geometry. However,
recreating this model in 3D in a GIS environment will be the more significant
challenge. The challenge begins at the onset and grows as we combine all
imaginary surfaces into a single model, while at the same time allowing for more
than one analysis to be run at a time. To meet this challenge, it will be critical to
know the existing capabilities of ArcGIS.
3.4 Geographic Information System
There are many improvements have been made in geographic information
systems during the last decade. They are very complex systems with extensive
capabilities, which are increasing from day to day as a growing number of
professionals discover their benefits. Currently, GIS is not just as a mapping tool,
but also as one of the primary analysis tools. As P. Van Oosterom states:
Geographic Information Science and cartography have been utilizing
GIS 3D models primarily for geo-visualization. Although visualization is
useful for obtaining visual insight and performing qualitative analysis, for
more applications a more quantitative analysis is needed.

In addition, GIS is capable of running 2D and 3D analyses and deriving
results from the complex database. ArcGIS can now provide services as spatial
analysis, visualization and cartography, and spatial data management. And as the
GIS community continues to demand more and more capabilities, the feature list
continues to extend.
However, despite huge analytical capabilities, GIS is currently unable to
provide the functionality needed for 3D volumetric modeling. True, we can
represent 3D models, using Arc Scene, and run 3D analyses in the ArcGIS
environment. It is possible to extend the current capabilities of any GIS system
through the use of scripts or the integration of customized modeling applications
with development of new modeling structures and extensive programming.
However, the current GIS platform itself does not support the creation and
presentation of a real 3D volumetric model.
Indeed, there are many more sophisticated and powerful programs, such as
3D Studio Max, Maya, Cinema 4D, and others with the capabilities and flexibility
of developing, presenting, and visualizing 3D models that far exceed the
capabilities of current GIS software. At the same time, these powerful 3D
software applications lack analytical capabilities and can only serve as a
representation tool or an environment for developing 3D applications that do not
provide support for spatial relationships.
There are many fields where a real 3D model is required over 2D and
2.5D models. For example, 3D models provide a clear representation of critical
data for medicine, archeology, and cadastre and airspace analyses, as 3D
representation increases the accuracy and capabilities of analyses that can be
preformed as well as facilitates the interpretation of results by presenting them in
more digestible form.
In summary, users have a number of methods available to them for 3D
modeling: through extensive software development, through the invention of new
modeling schemas, through maximizing the use of existing GIS capabilities, or
through development of specific procedures for creating a 3D model.
3.4.1 GIS concepts
To reiterate, the goal of this project is essentially to find the most effective
and efficient way for creating and using a specific pseudo 3D model within a GIS
environment. Presently GIS support three ways of surface models: Raster, TIN
and Terrain. All models can be used for Spatial or 3D analyses.

A raster is a geographic dataset in which values are assigned to a
rectangular array of objects'' (M. Goodchild, NCGIA).
Raster modeling is one of the most widely used methods in GIS. For
instance, the digital elevation model (DEM) is a raster model that is cell-based in
which each cell has a unique Z value. This model is optimally used when
representing data, such as rainfall, temperature, or concentration of chemicals.
However, it presents a problem when used in strict models where boundaries need
to be accurately defined as the raster data set has the edge effect and can store
more than one elevation for a single X, Y location.
On the other hand, there are many advantages to using this model
especially if you consider employing Spatial Analyst and Map Algebra. Spatial
Analyst allows for surface modeling, terrain, sophisticated raster data analysis,
and others. Another advantage is that raster data can be created from almost any
source of data, such as images, points, polylines, CAD drawings, etc. However,
due to the edge effect, the raster model is not suitable for developing an imaginary
surface model. Because the dimensions of the single cell describe the accuracy of
the raster model, this is another factor to take into consideration when building a
raster surface model. If the cell size were 10ft X 10ft, it would be sufficient for
land zoning or pollution analyses, but would fall short of the mark when
determining the perimeter of the imaginary surface.
Also, the Z value is valid for the entire cell. Therefore, if the cell size is
big, the accuracy becomes less so. On the other hand, if the cell size is small, the
accuracy remains high, but it also increases the size of the file and extends the
computing time. However, this model can be used for sophisticated surface
analyses if the imaginary surface model is created in another format, such as TIN,
and converted to a high-density raster surface.
Triangulated Irregular Network
The Triangulated Irregular Network (TIN) model is a vector-based digital
terrain model (DTM) which consists of contiguous non-overlapping triangles that
satisfy the Delaunay triangulation theory. Thus, a circle circumscribing a triangle
does not contain more points from the dataset in its interior. This 2.5D method is
used most appropriately in the representation of surface models and is not a real
3D volumetric model. However, three-dimensional data can be derived and
analyzed using a TIN model.
Using TIN, a constructed surface can hold only one Z value for one X, Y
location; however, a 3D Analyst in Arc Scene can represent a virtual 3D model
based on a TIN surface. The TIN model can be extended with the implementation

of the TIN/TEN structure as described by Verbree, E. and Oosterom, P. [8] in
their research on better surface representation by Delaunay Tetrahedronized
Irregular Network. This combination extends the capability of the existing TIN
structure, but requires that a completely new type of database be developed and
integrated with the Oracle Spatial database.
Presently, ArcGIS allows many functions and analyses to be preformed
with the TIN surface. It is also possible to convert TIN to raster, and raster to
TIN. As a result, there will be some loss of data, but we can consider it to be a
negligible loss. Thus, if we accept the conversion from TIN to raster and the
reverse process as necessary steps, we can, therefore, combine the usage of 3D
Analyst and Spatial Analyst and provide a solution that is powerful and it can run
almost any of the required types of analyses.
TIN surface can be visualized as 3D using Extrusion by knowing the base
height or using another TIN surface as its base. Volumetric calculations as cross
sectional areas and volumes are also supported if two or more TIN surface are
Terrain Model
The Terrain Surface model is by nature a TIN-based surface. It is built
from the point data stored as features in GIS database. Such sources are LIDAR
and HDS scanners. The terrain model allows pyramid levels to be created for a
particular display or analysis function with the use of two types of filtering:
Window and Z-filtering. The Window method filters data by selection and creates
pyramid layers by minimum, maximum, or mean values. The Z-filtering method
segregates data into different levels based on the specified rules applied to the
elevations. A key advantage that this model addresses is that the model is
automatically rebuilt each time new data is added. This advantage also meets one
of the essential requirements for airspace analysis and could be further developed
to address What If-scenarios.
All the previously mentioned surface models supported by GIS are not
true 3D volumetric models but they contain 3D data and can be presented in a 3D
environment. The main advantage of the existing GIS surface models is the
possibility of converting data between formats that, in turn, enable many different
types of analyses.
To build a real 3D representation for the purpose of airspace analysis, we
have to separate the features into three categories: existing terrain, restrictive
imaginary surfaces, and man-made objects. Each of these surfaces can be built
using raster, TIN or terrain modeling, but a final determination will be made
based on the existing data. As every type of surface has its own advantages and

disadvantages, it is good to use all of them to achieve more flexibility and expand
the analysis functions as well as work with the smallest file size.
For example, TIN and Terrain surface models are excellent for developing
and presenting natural surface phenomena, but the raster model allows for
extensive and more complicated analyses that involve the raster mathematical
toolset. Conversely, the raster model needs more storage space and increases, as
the cell size decreases. The smaller the cell size, the better the accuracy, however,
with a smaller cell size, the file size can increase dramatically. Thus, we have to
be very careful and take this into consideration as the computational time grows
with the files growth. However, in spite of that fact, I am confident that the
procedure proposed in this paper for developing the models and running airspace
analysis exclusively in GIS saves significant time and effort.
3.4.2 Spatial databases
The development of the PART77 model is related to a detailed
understanding of the different types of databases and their capabilities. To
determine which type of database is most suitable for this type of development
and serve as an appropriate repository for the data, it is needed to determine some
basic parameters. These include:
1. Type of basic file structure: shape file or feature class
2. If the file structure is a feature class, whether to use a Personal or
3. Capacity of the file structure
4. Support capability and the processing of Z values
5. If this environment, regardless of database type, could support all types of
primitives necessary for the creation of PART77 model
6. Flexibility of the structure itself to allow for adding new data or making
data updates
7. If this type of database or file allows for automation of the process
The selection of the appropriate type of database, datasets or files, which
store data and provide flexibility for querying, calculations, and process
automation are all important components for achieving the optimal solution for
PART77 GIS database development. In this case, GIS provides variety of choices:
Personal Database, File Database, and shape files with each type having its
limitations and strengths. The settings of the database for supporting Z values as
well as spatial accuracy are also two of the key factors for choosing GIS as

inappropriate settings and database selection will influence the whole modeling
process and can lead to unsatisfactory results.
A shape file is a single file structure, which can hold primitives like point,
line, and polygon. Further, it allows for processing and automation of the process.
Personal databases and geodatabase are structures that are far more
complicated, as they contain not only feature classes, but also feature datasets,
raster catalog data, and tables. The potential for the support and creation of
topology and topology rules are also not to be ignored, even though in the current
context, they are not widely used or needed. The selection between Personal
database and geodatabase is not very difficult because both structures are
seemingly identical and offer similar capabilities. Their distinguishing
characteristic is their size. Both MDB and GDB hold a huge amount of spatial
data with MDB holding usually no more than 500MB, and GDB holding up to
1TB. However, based on my preliminary calculations, the expected size of the
database for one runway will not exceed 50MB, which allows for the use of either
type of database.
Flexibility and capability for process automation are also significant
criteria that need to be observed for the success of the project. And specifically,
all three structures allow for optimization and automation of the process.
Therefore, in the beginning of the development of PART77 imaginary
surface model, I tested all three environments, and my results showed that all
three structures (shape, Personal, and geodatabase), can be used. The initial data
from geodetic measurements was presented as shape files and employed
throughout the creation of the PART77 model, after which, the end results were
moved and organized in the geodatabase.
3.4.3 GIS for 3D
From all the modeling schemas, I have described in this section, only one
model stands out as suitable for use in GIS at the present time. This model is the
Triangulated Irregular Network model and can be implemented immediately,
while the remaining models could not be implemented without the development
and integration of new software modules and new models of spatial relationship
in ArcGIS. The most promising of the group are TEN or the combined TIN/TEN
model, and currently the latter cannot be used as it relies on the specific use of an
Oracle Spatial database, and the latest Oracle llg has made tremendous
improvement in supporting and processing 3D models. As this type of model
seems to be potentially one of the best for 3D airport modeling and running

airspace analysis, I am planning to explore it further in future research regarding
the design of planning tools for airport planners.
3.5 High resolution measurements and data capture methods
Analyzing the current methods and practices for survey data collection
clearly shows that no single method stands out above the others as the most
effective and productive method for providing data for modeling three-
dimensional structures for airspace analysis. While one method is quick to
execute, it results in insufficient accuracy. Another method provides detail, but is
costly and time-consuming to collect data. Presently we can obtain spatial data
from many sources like Satellite imagery, Othophotos, GPS, LIDAR, HDS and
traditional survey methods (TSM) utilizing Total Station and Level instrument.
Each method has its own advantage and disadvantages and it is good for one or
another application. Depending on the accuracy we would like to achieve, we can
use one or more ways of data collection and even combine data to achieve our
goals. However, it is necessary to distinguish between horizontal and vertical
accuracy, and most of the methods while providing good horizontal accuracy also
provide poor vertical accuracy. To do justice to a complete discussion on data
collection methods and technology would fill volumes and extends beyond the
scope of this research project; however, I would like to provide a brief treatment
of the different types.
3.5.1 Imagery
Imagery is one of the most expeditious methods for collecting data in
contract to all other methods, as large areas of data can be captured within a few
hours or even minutes. There are two main types of images: satellite and
photogrammetric, and both types can produce an enormous quantity of
information. Images can be panchromatic, color, multi spectral and IR. The
satellite image horizontal accuracy is around lm, but Geoeye -1 provides 0.41m
for a BW image.
The photogrammetric data collection method, which uses a digital or film
camera mounted on an airplane, is more accurate than satellite imaging, primarily
due to the altitude of aircraft. The larger the photo scale (lower the aircraft), the

more accurate the data tends to be. Additional factors as camera type, operator
expertise, and ground control density play significant roles, and depending on the
project area and project purpose, the vertical accuracy can be up to 10cm.
Therefore, image data can be used for many applications and for the
creation of an accurate Digital Terrain Model (DTM). While images have a huge
advantage for providing accurate data for large areas and for DTM creation or
manmade structure database development, at the same time, they pose several
significant disadvantages. Images cannot provide data inside buildings, do not
provide good data for the existing ground in a forested area, building shadow
create a challenge and require a post-processing period before the image is ready
for use.
3.5.2 LIDAR and HDS
Modem airborne LIDAR terrain mapping provides a 0.15-0.3m-elevation
accuracy, depending on the terrain and mission parameters, and is cost-effective
for facility level applications [16]. Horizontal accuracy can be as good as 70cm.
Currently, it is one of the newest and most accurate methods available but there is
no doubt that by far the method that outweighs all others in regard to accuracy,
remains to be conventional ground surveying techniques, with the most advanced
methods utilizing High Definition Surveying (HDS) scanners. However, this
method is time-consuming and costly to execute and collect data. For this reason,
many researches have disregarded HDS as a viable solution for data collection in
airspace modeling.
Although both LIDAR and HDS use laser scan technology, the main
difference between the two methods is the accuracy they provide and the area
they can cover within a specified period of time. LIDAR is a technology that
provides large datasets for large areas, but it also requires extensive follow-on
time for data cleaning and interpretation before the date can be used for DTM
creation. Further, the LIDAR method provides a good source of data for terrain
and manmade structure modeling, and there are inexpensive off-the-shelf software
packages on the market that can be used to work with LIDAR data. GIS tool
LAS to Multipoint allows for LIDAR data conversion into feature classes, and
once data is converted to a multipoint feature class, the data can be reused for
building other features, 3D presentation or analyses.
At present, HDS is the most accurate method available, able to provide up
to 2mm of positional and elevation accuracy, but it is time consuming to
implement. The latest LEICA Scan Station CIO can scan up to 300m with a 4mm

positional accuracy and create DTM with a 2mm accuracy. HDS provides highly
accurate data for 3D modeling of open terrain, as well as inside buildings, and can
capture color images that can be overlaid and used for rendering of the model.
This instrument also has the capacity to solve a series of surveying problems in
the field and is compatible with standard surveying equipment and the GPS unit
for station positioning.
3.5.3 GPS
GPS technology is now widely used every day in survey data collection
and layouts. Currently, GPS is used almost in every construction or surveying
project. This technology has grown rapidly during the last 10-15 years and has
become one of the most used in the world. This is due in great part to the fact that
a GPS has become more and more accurate. The latest development made by
JAVAD and based on their Triumph chip is capable of tracking 216 channels and
all GNSS satellites provide a RTK accuracy of lcm+lppm horizontally and
1.5cm+1.5ppm vertically. This is truly tremendous progress.
Presently, GPS is being developed more and more for all types of
measurements. However, a single disadvantage to the GPS is that it cannot be
used for accurately measuring inside buildings, over forested areas, and within
densely structured urban areas due to multipath and ambiguity of the signal in
these areas.
3.5.4 Traditional ground survey methods
The traditional ground survey methods are still one of the most accurate
ways for data collection. They provide a flexible means for gathering different
data in the field, inside buildings and virtually everywhere. There is almost no
restriction for the use of this method in data collection. With the new Total
Stations and digital Level instruments, the accuracy of data is almost unbeatable
by any other device, excluding, of course, HDS. Its disadvantages are that it is
very time consuming to implement, requires existing ground survey control points
to serve as a reference framework and visibility between the points. Another
shortcoming is that we can use a traditional method of data collection only during

daylight hours. However, despite these problems, this type of survey provides an
immense source of highly accurate 3D data, using the most convenient units -
In conclusion, I can safely say that all type of surveys can be used in
supplying necessary data for the development of 3D models, as all methods
described in this section provide three-dimensional data. As to which is the best
method to apply to a specific type of model development, an airport planner must
determine which method ensures meeting the targeted model accuracy, which
falls within the projects financial budget, and which provide sufficient support
and data availability.

4. Proposed airspace analysis model
4.1 Overview
In the Chapter 2 of this paper, I discussed different types of airspace
analyses. However, I would now like to focus on preparing a PART77 model that
will be used in specific airport runway airspace analyses (ARAA). As stated
earlier, the most important ARAA for airport planners are LOS-ATCT, TERPS,
and HDS with multiple ways of running each of these analyses. Each method
depends on the interpretation of FAA documentation related to them, the data and
software availability, the goals of the analysis, and whether the analysis operates
as a single-point analysis or what-if scenario for future development.
During the time I worked as consultant for DIA, I was in constant
communication with other planners from other commercial airports. Through our
shared experiences, I was able to learn what type of data they used, and how they
conducted ARAA. It was during that process that I realized that most airports do
not share a consistent means for conducting ARAA analyses, and, further, do not
use pre-built models to run ARAA. Because of this, I ran a number of different
types of analyses based on different requirements and scenarios, and at the same
time, was able to gain experience on the process, while collecting information
about ARAA.
Further, during the process it became clear to me that some methods
surfaced as being more effective than others. These included the use of GIS, the
process of building spatial models and frequently updating them, and streamlining
the development procedures for conducting any one of the aforementioned ARAA
analysis types. It was a culmination of these efforts, including some additional
investigations and research that resulted in development of the proposed airport
airspace model in GIS and the streamlined and automated process for
development of PART77 that I will discuss in this chapter.

4.2 PART77 and ARAA as general processes in CAD
The process of developing a full PART77 imaginary surface map for a single
runway inside a CAD environment requires many man-hours and involves
numerous types of calculations. As most calculations need to be made outside the
CAD environment and the resulting computed data entered back into the CAD
system, this scenario lays the groundwork for a host of human errors, particularly
when entering data. If a particular airport has more than one runway, which is
common for commercial airports, enormous hours of hand-made calculations
could be required as well as the need for a combined imaginary surface map for
the entire airport. The latter operation in-tum would also involve numerous
mathematical calculations and many additional hours of work, with a timetable
approaching upwards of several days to complete a full airport PART77 map.
Further, as the current CAD environment is not sufficiently sophisticated to
support a database, imaginary surfaces designs for one PART77 runway would
typically need to be constructed as separate layers. See Figure 10, where the
numerical values describe the number of surfaces.
(based ora runway end points)

The PART77 map in CAD is presented as 2D drawings. To make it suitable for
ARAA it requires 3D polygons to be developed. Later each polygons should be
used to create surface.
Having each runway as a separate drawing means there are imaginary
surfaces only for this particular runway, and the drawing file is relatively small.
However, it is difficult to run an analysis, which involves more than one runway
for the PART77 model. Yet, if all surface for all runways are developed in the
same drawing, the results are a bigger file and more surfaces (For example: 3
runways X 7 surfaces = 21 surfaces) and compounds the possibility of human
error. The large size of the file also slows down the process of analyses and a
simple analysis in CAD could take an inordinately long time. Figure 11 below
shows this workflow for an Airspace Analysis in a CAD environment.

* ORTmeeDFFErtexTiuMnATSfKm
possible soumoNsmomMs

From Figure 11, it is easy to see that all the manipulations are manually driven, as
there is very little possibility to automate such a complex process in a CAD
Therefore, to remedy this situation, I propose developing a GIS database structure
that will hold the data for PART77 and a methodology to create the PART77
model, a model that I have been using and constructing for more than one year
while consulting at DIA. The details of this model and its methodology will be
described in full in Chapter 6 of this paper.
4.3 FAA specifications
The first step in my development process was to thoroughly review all the FAA
documentation related to PART77. These included:
Title 14 Code of Federal Regulations (CFR) PART77 Objects Affecting
Navigable Airspace
This document describes each of the imaginary surfaces, provides key as
well as information for developing PART77 for one runwayone runway
consisting of Primary, Horizontal, Conical, Transitional and Approach
imaginary surfaces, parameters and gives information as to how surfaces
can relate to one another.
United States Standard for Terminal Instrument Procedures (TERPS) -
FAA Order 8260.3B
Further, contained within these documents is a requirement that mandates that if
there is more than one runway, all runway PART77 models need to be combined
to create an airport PART77 model. However, the FAA specification does not lay
out clear guidelines to how to conduct this process.
4.4 3D Mathematical foundations
The second step in my development process was to formulate a mathematical
foundation for all the calculations that I would need in building the PART77
model. Note that because the PART77 model is a combination of 3D imaginary
surfaces, it consequently requires calculations in three dimensions (X, Y, Z) to

determine the position and elevation of each imaginary surface. Surfaces can be
described by the surface equation, 3D surface boundary, or a series of points.
Developing an equation for each particular surface is a good first approach, but
there are some notable constraints with this method. First, working in GIS and
using surface equations would normally lead to extensive programming, which
fell outside my project goals. Second, because each runway surface is somehow
unique, there could be no possible means of covering all types of surfaces with
each surface using its own equation. Third, while investigating this method, I
realized that one of the best means for achieving my goal was through the use of
control points [38] and [39], But, then again, if I already knew the control points
with their 3D coordinates, there would be no need to construct an equation
because once I knew the control points, I could easily build the surface
boundaries. Further, I do not know all coordinates for each end of the surface
boundary. And, moreover, some surface intersection lines are presented as a series
of points. Consequently, I would have had to know the position of these points in
order to find the way to calculate their 3D coordinates.
Thus, the model I finally developed was based on significant points
(SIGPOINTS) determined as 3D positional coordinates in space. The relationship
between the position and the elevation is quite simple, but every boundary line is
dependent on its adjacent surface and must follow a specific azimuth, which does
not allow the usage of a normal linear equation. The surface boundary line should
be also presented as 3D, and not as 2D, but the calculation of dx, dy and dz is
based on the horizontal distance rather than the slope distance. So, using a
parametric equation for the 3D line would not solve the problem, as they do not
involve azimuths.
Therefore, I employed the use of parametric forms of general equations
that are typically used in surveying, that is, two equations to determine the
position and one equation for calculating the elevation of these points. The
parametric forms of these equations are as follows:
[1] Xn = Xn-i + (HD sin (a))
[2] Yn = Yn-i + (HD sin (a))
[3] Zn = Zn-i + (HD* m)
where Xn, Yn, Zn are the coordinates and the elevation of the relevant runway
centerline point or preceding point of this particular boundary line,
HD is the horizontal distance between two points,
a is the azimuth,
m is the slope of the line.

In the case where the distance is not known, but there is a known or required
elevation(s) and preset slope angle, the solution for distance would take this form:
[4] HD = (Zr -Zn) / tan (y)
where Zr is the required elevation,
Zn is the elevation of starting point,
and y is the slope angle.
At first glance, this method appears simple to implement, but in reality, it is not.
Normally in surveying, measured distances and azimuth are known elements.
Here they are not, and consequently, must be derived from runway centerline
points and other supplemental calculations. So, to meet this requirement, a
parameter file was created to hold all the given parameters. I also developed the
equations to compute the parameters that were not given and needed to be
calculated, but were absolutely necessary to complete the SIGPOINTS
computation. The parameters presented in FAA documents were as follows:
(Values presented here are valid for only one specific runway)
Primary surface Extend
Primary surface width
Clear way
Airport Elevation
Horizontal surface radius
Conical surface distance
Conical surface slope 20:1
Outer transitional surface width
Transitional surface slope 7:1
Approach surface 40:1 end width
PSE = 200ft
PSW = 500ft
CW = Oft
AE = 5434ft
HSR = 10000ft
CSD = 4000ft
CSS =0.05
OTSW = 5000ft
TSS =0.142857
ASEW = 16000ft

Approach surface 40:1 Length
AS401L = 50000ft
Approach surface 40:1 slope
AS401 =0.025
Approach surface 50:1 end width
AS501EW = 4000ft
Approach surface 50:1 length
AS501L = 10000ft
Approach surface 50:1 slope
AS501 =0.02
The calculation of missing parameters and the equation used to compute them:
Horizontal Surface Elevation (HSEL)
[5] HSEL = AE + 150 (ft)
Conical Surface Elevation (CSEL)
[6] CSEL= HSEL + (CSS*CSD) (ft)
Width of Approach Surface 50:1 at the end of 10000ft
This parameter is not provided in the FAA document but is a required calculation
in coordinating the computations for significant points.
Calculate half distance at 401 end width
[7] dil = (AS401EW- (2 PSW)) / 2 (ft)
Calculate half distance for 501 end width
[8] di2 = (dil AS501L) /AS401L (ft)
Calculate 501 end width
[9] AS501EW = (di2) *2 + PSW 2 (ft)
The deviation for the Approach surface points from the Primary Surface point can
be presented as Distance and Angle.
Calculate distance from PS point to 50:1 point
[10] DItoSOl = Sqr((AS501L A 2) + (di2 A 2)) (ft)
Calculate distance from PS point to 40:1 point
[11] DIto401 = Sqr((AS401L A 2) + (dil A 2)) DIto501 = DIto501 (ft)
[12] DIto401 = DIto401 (ft)

Angle of deviation from CLRW Azimuth
[13] AngAS = Atn(dil /AS401L) (radians)
This final computation results appear in Table 1 showing the Distance to and
Deviation Angle from the CLRW for all types of runway.
Extend Width Distance ANGLE
0 250 250 0
0 500 500 0
0 1000 1000 0
200 250 320.1562 51.34019
200 500 538.5165 68.19859
In conclusion, these parameters provided enough data to compute the X, Y, and Z
coordinates of SIGPOINTS. There are thirty-two significant points that
determined the full description of the PART77 surface for each runway, and these
points had to be calculated and connected with 3D polylines in order to describe
each imaginary surface. The 3D lines distinguished between different surfaces
and served as a boundary for each surface.
All points fell under one of the following categories:
Primary Surface points which describe the end of primary surface (PSP)
Inner Transitional Surface points which describe where the inner
transitional surface meets the Horizontal surface (ITSP). To present this
line accurately as an intersection line (IL), it is also necessary to calculate
these points coordinates as offsets of centerline points.
Conical Surface points are the points that describe the intersection of the
Inner Transitional Surface and the Conical Surface (CSP)
Outer Transitional Surface points describe the limit of the Outer
Transitional Surface (OTSP)

Approach Surface 50:1 points represent the end of the 50:1 approach
surface at a length of 10000 ft. (AS501P)
Approach Surface 40:1 points determine the end of this surface at 50000
ft. extending from the primary surface to a width of 16000 ft. (AS401P)
The calculations of SIGPOINTS coordinates were set as a Visual Basic (VB)
script, utilizing all the equations presented in this section, and the resulting files
and the GIS capabilities were used to develop the PART77 model.
4.5 GIS methodology and 3D spatial representation
GIS provided a wide range of tools and methods for the database
development and data processing portions of the project. As indicated earlier, the
use of these capabilities was central to the goal of my research, as I specifically
did not want to include the use of any third-party software or separate and
extensive software development.
However, the data provided by the FAA documentation and the results
from the calculations I laid out in the previous section cannot, in and of
themselves, provide the means necessary to create the model I propose.
Additionally, the process involved a specific methodology that I developed in
order to achieve the goals of developing an effective model for airspace analysis,
while streamlining the procedure to ensure a minimum of direct human
In summary, the methodology involved incorporating the VB script for
mathematical calculations, the data management tools for the database
development, and a toolbox I developed to streamline the entire process. I
sequenced these steps in a defined and prescribed order that start with entering of
data and culminate with the building of a final, complete, and cohesive model,
with the presentation of the final data appearing in 3D spatial form, using Arc
Scene to create the 3D-representations of the model and analysis results.

5. Data collection / acquisition:
To assure correct and accurate results for airport runway airspace
analyses, highly accurate survey data are required in the model development.
Three models were identified, PART77, EMSM and ESGM to be part of ARAA.
Because each of the models consists of different features, each will have different
accuracy standards. These standards should follow or exceed current FAA
requirements. The latest FAA AC 150/5300-18B determines the accuracy level
for surveying the airport features from the prospective of building eALP in a GIS
environment, and although most of requirements for data accuracy are good, in
some cases standards are too broad to be effective when building the models.
5.1 Existing Ground surface model
Existing ground surface models are usually presented as contours and spot
elevations. All sources of data are acceptable to build this model. Usually
photogrammetry or LIDAR measurements supply the data for developing the
contour map and the existing ground surface. Small contour spacing means
greater accuracy and better representation of the existing ground surface, and with
the use of a contour map, it is possible to create another surface with denser
contours. This is purely a mathematical process involving interpolation and does
not in any way ensure that the created surface is more correct.
Generally, the original data is based on 1ft or 2ft contours. I consider 1ft
contours to be an acceptable level of accuracy for EGSM as collecting more
precise data for a large area is not only quite difficult, it is often also cost-

5.2 Existing manmade structures model
Existing manmade structure models represent all artificial objects, such as
buildings, towers, antennas, bridges, and the modeling process of these structures
is still very difficult in a GIS environment, as ArcGIS does not support full 3D
models for such structures. Data for these features can come from ground surveys,
LIDAR, HDS, or photogrammetry. Again, all data sources are acceptable. The
choice of data source depends on how accurately we want to build the model.
FAA 150/5300-18B specifies 1 ft accuracy in the vertical and horizontal planes for
all buildings, and based on my experience, this is a satisfactory standard. Models
for buildings can be also created from as-build drawings. In the latter case, the
models will be much more accurate, as these drawings are drafted with a great
degree of engineering accuracy.
5.3 PART77 Model
While all models require accuracy, the PART77 model needs to be the
more accurate of the two. As mentioned earlier, generally only runway end points
are used for PART77 development. However, a closer examination reveals that
this is not the best approach, as two points can be used only if the runway is
completely straight, and there are not vertical curves. The presence of even a
single slight vertical curve can make the model inaccurate, so the most affected
surface in this case would be a Transitional surface 7:1.
A surface extending outward and upward at right angles to the runway
centerline and the runway centerline extended, from the sides of the
Primary surface and the Approach surface: the slope is 7:1 and the
surface extends until it intersects the Horizontal or Conical surface. [30]
Thus, incorrect placement of the intersection line will narrow or widen the
Transitional surface.
The distance of the point of intersection between the Transitional and
Horizontal surfaces is a function of existing centerline point elevation (El) and
Horizontal Surface Elevation (HSE), as shown in the following equation, where
0.143 represents 7.1 Transitional surface slope.
[14] Distance = (HSE El) 70.143

To clarify this, I calculated the variance between two models for
Transitional 7:1 surface: one model based on two points, and another model
based on the centerline points. The result is presented in Figure 12. Here, you can
see that a lateral difference of 476ft exists between the two models: one based on
runway end points, and another based on the centerline points, resulting in a huge
impact on any planners decision. Therefore, the use of centerline points will
definitely increase the models accuracy.

Additionally, the centerline points themselves need to be accurately
surveyed as well. Consider, for example, that the farthest point of any PART77
surface is 50000ft away from the runway end. The difference of 1ft horizontally

and/or vertically will result in large vertical and horizontal deviations from the
correct surface position. Therefore, more stringently accurate centerline point
settings are required as these points will be used as base data for developing the
required PART77 imaginary surfaces, as described in the FAA FAR49 PART77
Runway centerline data should always be collected using a GPS or Total
Station, and, if possible, a High Definition Survey scanner. All data must be
contained within one coordinated system, utilizing predetermined vertical and
horizontal datums. Some airports, such as Denver International Airport (DIA),
have a local coordinate system. However, apart from airport construction,
working with a local system provides little advantage from the prospective of
eALP and AGIS. Further, if a local coordinate system is used, it will be difficult
to present the results to outside organizations and stakeholders, who do not in turn
utilize this same system. And, in addition to data sharing and data presentation,
further concerns can be raised based on how accurately and how well defined the
transformation is among the local system, the State Plane Coordinate System, and
the Geographic Coordinate System.
For purposes of this research and to avoid any potential problems, I used
SPCS NAD83 CO CZ FIPS 0502 (ft) as the main coordinate system and a vertical
datum of NAVD 88 (ft) with a horizontal datum of NAD83. Data for the
centerline runway was collected, using a GPS with a 0.2ft vertical and horizontal
accuracy, and each runway was represented as a series of points in a simple
comma-delimited format file (CLRWname.csv) as shown in Figure 13

C3 Microsoft Excel RWYlf>R3-*lspcsxsv
.HglJ Bte Edit Mow Insert Format loots Data JMndow ttatp Adobe PDF
'i. V3 U Li (iJ;a lAJ.^JSU * .iJi MlilbJi
117 V fit
1 'B = iISr'a is. D a' s -
Ai, PH X Y z Name
E2li 2513 3225571.76 1752178.84 5320.55 RWY END RWY 16R NAVD88
3 2543 3225571.76 1752112.19 5317.53 RWY 16R-34L CL (NAVD88)
2578 3225571.77 1752028 88 5317.66 RWY 16R-34L CL (NAVD88)
rs 2610 3225571 77 1751945.57 5317.78 RWY 16R-34L CL (NAVD88)
2637 3225571.77 1751866.92 5317.90 RWY 16R-34L CL (NAVD88)
*r-i 2669 3225571.77 1751779.44 5318.05 RWY 16R-34L CL CNAVD88)
sm 2699 3225571 77 1751689.46 5318.19 RWY 16R-34L CL fNAVDSSI
W 2734 3225571 77 1751599 48 5318.33 RWY 16R-34L CL (NAVD881
2768 3225571.77 1751509.50 5318.47 RWY 16R-34L CL (NAVD88)
urn 2803 3225571.78 1751419.52 5318.63 RWY 16R-34L CL (NAVD881
?12l 2838 3225571.78 1751329 54 5318.79 RWY 16R-34L CL (NAVD88)
-'13; 2874 3225571.78 1751239.56 531888 RWY 1BR-34L CL (NAVD08)
2909 3225571 78 1751149 58 5319.02 RWY 16R-34L CL (NAVD88)
2944 3225571.78 1751059.60 5319.16 RWY 16R-34L CL (NAVD88)
16 2980 3225571.78 1750969 62 5319.34 RWY 16R-34L CL fNAVD68)
Ihr 3060 3225571 78 1750879.65 5319.46 RWY 16R-34L CL fNAVDOS)
MSI 3089 3225571.78 1750789.67 5319.61 RWY 16R-34L CL (NAVD88)
19 3151 3225571.79 1750699.69 5319.74 RWY 16R-34L CL (NAVD881
$0 3202 3225571.79 1750609 71 5319.83 RWY 16R-34L CL (NAVD68)
tzm 3304 3225571.79 1750429.75 5320.13 RWY 16R-34L CL (NAVD88)
:-2Z> 3340 3225571.79 1750339 77 5320.25 RWY 16R-34L CL (NAVD881
3375 3225571.79 1750249.79 5320.39 RWY 16R-34L CL (NAVD68)
m 3411 3225571.79 1750159 81 5320.53 RWY 16R-34L CL CNAVD88)
*2tt 3446 3225571.79 1750069.83 5320.70 RWY 16R-34L CL (NAVD881
3481 3225571.80 1749979.85 5320.82 RWY 16R-34L CL 1NAVD88)
3517 3225571.80 1749889 87 5320.94 RWY 16R-34L CL (NAVD881
3552 3225571 80 1749799 89 5321.06 RWY 16R-34L CL (NAVD881
m 3587 3225571.80 1749709.92 5321.23 RWY 16R-34L CL (NAVD68)
Where PN is the point number, X is the East coordinate, Y is the North
coordinate, Z is the Elevation, and Name is the point name or description (in this
case it identifies the runway designation). This field also can identify the station
number along the runway centerline.

6. GIS based airspace analysis model
6.1 Overview
To determine the best methods for development, and the mathematical
equations and logical procedures that needed to be used, I first needed to ascertain
whether it was necessary to build a full PART77 airport model that would be
employed across all types of analyses, or build individual models for each and
every surface and later combine these surface models as needed to create the
relevant analysis.
As a result of this process, four possible ways were identified and
investigated. These are presented in Figure 14.

As illustrated in Figure 14, the model development process can be
separated by data type:
Development from scratch using initial surveying data
Development from existing CAD drawings
or by model type:
Whole Airport PART77 model
Individual PART77 surface models
6.1.1 Variant 1 Full PART77 model
A Full PART77 model includes all surfaces from all runways, resulting in
a very complete, yet complex model. Typically, it represents the most restrictive
parts of each surface as a combination of all imaginary surfaces. The question that
arises with the use of this model is: Is it possible to create this model without
calculating every surface separately beforehand?
FAA documentation provides a description of each and every surface, and
each surface has distinct parameters based on the runway type. Moreover, all
surfaces are adjacent to each other and are dependent on each other as they share
common boundaries. Hence, without creating a model for each imaginary surface
separately beforehand, the creation of the whole model is not feasible.
6.1.2 Variant 2 Combined individual PART77 surface models to form a
Full PART77 model
A lot of time is required to develop the logical structures and mathematical
equations that describe every airport surface, as well as to develop a defined
procedure that ensures that the entire process yields an accurate and correct Full
PART77 Runway model.
On the positive side, we have data for each and every surface and runway,
enabling the manipulation of those individual parts and their combination with the
other surfaces, as needed. This factor underscores the flexibility and pliability of

the model itself. Also, after the mathematical equations and the logical structure
of every surface and the links between each of the surfaces are established, further
work or rework is made easier, as the equations and processes can be reused for
the creation of new surfaces types, simply with different beginning parametrics.
Thus, we can conclude that:
1. Variant 1 is impossible without creation of every separate surface,
which automatically leads to Variant 2.
2. To complete Variant 2, the following steps must be followed:
a. Find the description of every surface by FAA document
b. Create a list of parameters and given variables
c. Create a list of the missing parameters and variables
d. Develop the mathematical equations needed for calculation
e. Develop a logical procedure for creation of the surface
f. Create an automated procedure for the implementation of
equations and logical procedures
g. Create a database for each runway
h. Create a database for the whole airport.
Briefly, by comparing the data type methods described earlier in this
paper, we can conclude that the development from scratch using survey data is a
more effective approach than using data from other sources. In the former
method, new and accurate data for the full length of the runway centerline is used.
This approach ensures accuracy, because the type of collected data can be preset
and the development process can not only be fully controlled but also rechecked,
if needed. On the other hand, as most airports already have a PART77 map in
CAD, it seems to be more reasonable to use the existing CAD data. However, it
must be kept in mind, that normally CAD drawings lack sufficiently accurate data
to provide an effective analysis and some data may be lost during the data
migration from CAD to GIS.
A layout of these two methodologies is illustrated in Figure 15.

Both methodologies and the steps involved to execute each are described
in detail in the following sections.
6.2 PART77 model development from existing CAD drawings
When using the PART77 model development from existing CAD
drawings, we are immediately faced with the fact that there is one drawing model
for each runway. Although it is not necessary to develop specific mathematical
descriptions for each surface as the surfaces already have been developed in the

CAD environment, it is necessary to find the right procedure to properly transfer
surface data into GIS environment as to alleviate any type of data loss.
The PART77 model for each runway consists of Primary, Horizontal,
Conical, Approach and Transitional surfaces. Approach surfaces are designed for
each end of the runway, so there are two parts for each approach surface per
runway model. The situation with Transitional surface is the same. There are two
transitional surfaces: one for each side of the runway, the left and the right, and
each side have two parts, the Inner and Outer Transitional surface. Therefore,
there are seven polygons in the CAD drawing that represent all surfaces, and each
surface is represented as a layer.
The diagram in Figure 16 shows the structure of all the surfaces and their

However, the Approach and Transitional surfaces are not really split on
four separate polygons (40:1 and 50:1 for Approach and Inner and Outer for
Transitional) as shown in the preceding figure. The diagram illustrates the surface

components; however, only one polygon represents each Transitional side and
each Approach side surface.
When the CAD data is imported into an Arc Map environment, there are
six layers that represent the different primitives: Annotation, Point, Polyline,
Polygon, Multipatch, plus the drawing layer that holds the symbology
information. The TIN surface can be created from points, polylines, polygons and
multipatch, or any combination therein. And depending on the different settings
for the data types, triangulation can be set as soft clip, hard clip, soft erase, or hard
erase, etc. Then, based on this setting, ArcGIS treats the data differently from the
future classes, and it is reflected on the correctness and accuracy of the model.
Further, it seems relatively easy to import the CAD file into an Arc Map and
export all layers as shape files, converting all features classes like Polygons or
Multipatch into shape files, and finally creating the TIN model from these
Another option is to create a TIN model, using surface boundary polylines
or contour polylines. The latter approach seemed to be the best, as it describes the
surface exactly as it was developed in CAD drawing; however, the resulting file
size could be very large, and which would in turn affect the size of TIN or Raster
PART77 surface model as well. Overall, the processing time depends on number
of features involved in calculations. The more features in file, the longer it took to
create the TIN surface. As a result, I ran several investigations into this problem
and tested several different scenarios. I also created a TIN model from different
set of features.
The table presented in Figure 17 shows a summary of the results of these
A Source TIN TIN size L....P-J Shape
Mil paid) cannot be created NA Milpatch
Polygons soft do. not ikW in some areas 35MB PotygonZ
Potyirtes -surface boundaries connections outside boundary, delineation isnotwork orooertv 1MB PoMneZ
PoWnes-contours harine, connections outside boundary, defineation works but needs several attempt; 247MB PolyfneZ
From Figure 17, we can see that none of the scenarios created correct TIN
surfaces that matched the PART77 model in the CAD drawing. And, in some
situations, even the use of the ArcToolbox Delineation tool did not help to
calculate the right models. Investigating the problem further, I found that the

elevation data, which is essential for terrain modeling, was missing or somehow
transferred incorrectly when the polygons or multipatch were exported. With
further research, I found that this occurred because the existing CAD fdes had
different settings for their polylines, and the behavior of exported polylines
differed from file to file. The result was that I had to come up with another
method other than using the data form polygons or multipatch.
I concluded that the only way to get right data from the CAD file and
create a shape file from which I could develop a correct TIN model was to use the
polylines that described the surface boundaries (SBP). The SBP did contain the
needed elevation data, but still did not create the right TIN file, and in the end, it
was required to edit the model.
Subsequently, I developed a TIN model from the polylines shape file
running 3D Analyst, but again the resulting TIN file was not completely right, and
I was forced to further investigate the problem and redefine my procedure. At that
point, I discovered that polylines do not act as hardedges, even though it was the
default setting for the triangulation. So, my next step was to use the Delineate
function in Arc Map and trim away everything that fell outside the desired
boundary. However, this method proved to be ineffective as it resulted in some
incorrect calculations within the TIN model. Additionally, it took several
iterations of this process as well as different delineation and triangulation settings
to calculate the best results, and the results were different for different CAD files.
On top of that, I discovered an additional problem when exporting data
from CAD into a shape file. In both cases, whether I used the CAD export
function or GIS export data function, the export file was excessively large, having
inherited all the information from the CAD layer. Most of this attribute data,
presented in different columns, held unnecessary information, and again needed to
be deleted. The only valuable information that I found needed to be kept was the
elevation data. Therefore, I created a tool in Model Builder that automatically
deleted all these columns after the data export was complete.
As a summary of my investigations, I concluded that:
1. Polygon and Multipatch data do not hold Z values when exported into a
shape file.
2. Because of factor #1, the TIN model cannot be built or, if built, is
incorrectly structured.
3. Contour polylines are exported correctly and can be used for a TIN model,
but the model is not correct and several iterations of the Delineation tool
and triangulation settings are needed to make the model correct. The
reiteration of the TIN model was not easy, but when I achieved the desired
model, the resulting model had a file size that was excessively large.

4. Polylines, describing surface boundaries (SBP), were transferred correctly
into the shape file, and the TIN model was created, but the model was not
consistently correct every time.
As one of my goals was to create a general procedure that could work with
any type of existing PART77 CAD file, I moved forward and refined the process
until I was able to provide consistent results, and the SBP were used to describe
imaginary surface boundaries.
So, the final workflow to build a PART77 GIS TIN model from an
existing PART77 CAD file, using surface boundary polylines, is presented in
Figure 18 here.

The process started with the selection of 3D surface boundary polylines
(SBP) from imaginary surface layers and the exportation of those lines to a
LINE.shp file. As I already indicated, the TIN model created from Polylines
always had some errors due to the incorrect treatment of polylines as not being
hard edges. To solve this problem I used a topology tool to construct polygon
topology that was constructed from the SBP file. Using the topology tool, a 3D

polygon feature class was developed, and this class inherited Z values from the
Polyline feature class, but restricted the TIN model creation to be inside the
polygon area only. Finally, the resulting polygon feature class consistently created
the right PART77 TIN model. Then, to allow for a faster, better and more flexible
way for airspace analyses, the PART77 TIN model was converted to a raster
model. The complete workflow as designed in Model Builder is presented in
Figure 19.
After the process was developed and the ensuing problems solved, the
next step in the process was to find out how and where to store the data for the
runway PART77 model. The directory and geodatabase structure for the
individual runway PART77 models developed from existing CAD drawings is
illustrated in Figure 20.

(Existing Drawing tile)
All the data transferred from the existing PART77 CAD drawing, TIN
models, and raster models were saved into separate subdirectories, and the
runway geodatabase holds the information for line and polygon feature classes. It
is important to interject that the PART77 model represents one runway model
only, and that the models for each subsequent imaginary surface could be easily
created from the relevant polygon or line feature class and saved as raster and TIN
models for future use.
This logical structure was chosen because it enabled the storage of all data
in a single repository, enabling easy access to the needed data store and providing
sufficient flexibility to allow for the adding of data as files, feature classes or TIN
and raster models, as needed. Examples from each data set, CAD, Lines,
Polygons, TIN and Raster are presented as illustrations in Appendix A CAD to
GIS Modeling.

I would also like to point out that Model Builder was used to streamline
and automate the process of transferring data from the existing PART77 CAD
drawings to the GIS database and that the PART77 TIN and Raster models were
built in a GIS environment. The tool was able to process one CAD file at a time or
run batch jobs when working with multiple CAD drawings at one time.
6.3 Development from scratch using initial surveying data
This method is far mode complicated than the CAD-to-GIS method. The
complexity arose from the fact that there was no previously developed model
from which to derive information, as well as no series of processes to lead to the
creation of the PART77 GIS model for every runway. However, despite those
drawbacks, this method proved to provide a much bigger advantage than CAD-to-
GIS method, because the initial surveying data provided more accurate and more
detailed data than those data taken from existing CAD files. As noted earlier, it is
not always known what data are used for the creation of CAD PART77 and how
accurate this data is, and that the existing PART77 models in CAD files are
created based on runway end points. Thus, the use of survey data that describes
the runway centerline led to a more consistent and reliable PART77 model.
To test this method, I began by collecting survey data for the model using
GPS. The horizontal and vertical accuracy was 0.2ft. with the Horizontal Datum
being NAD83, Vertical Datum being NAVD88, and all points remaining within
the State Plane Coordinate System, Colorado Central Zone FIPS 0502 Feet.
Then, to build every surface using the survey data as a base, I needed to
know the relevant surface geometry, taken from a description of the FAA
PART77 documentation, and the links between each surface to determine the
exact order of their creation. The table I used from the PART77 documentation is
illustrated in Figure 1. Here, we can see that some of the basic parameters of the
Primary, Horizontal, and Approach surfaces depend on the type of runway. That
is to say, if the given runway type is a Utility and also a Visual approach
runway, the width of the Primary surface would be 250ft; but if the runway types
are Utility and Precision Instrument Runway, the width of the Primary
Surface would be 1000ft. These parameters, taken from the FAA documentation,
plus their relationship to every imaginary surface, are shown in Figure 21.




V~x jf The slope b 7*1 (143 percent) Had ihe surface extends uatll'i

/: HK ApproachSurfecettutproJcct beyanddiellnilbofthf \j
Conical Surface extends a distance of 5,009 float ensured
horizontally from the edge of Ibc Approach harficc Ihc vJopc la
1-1 (14J pcimi]. j *


34:1 -1
*i, ^ I
f mm

In addition to the previously mentioned parameters taken from the FAA
documentation, there are three additional parameters that affect the creation of
imaginary surfaces and are given by default. These parameters, Airport Elevation
(AE), Horizontal Surface Elevation (HSEL), and Conical Surface Elevation
(CSEL), are listed in the top-left purple box in Figure 22

The last two parameters of the group, HSEL and CSEL, are functions of
the Airport Elevation that are by the FAA for each airport. As most of these
parameters are known, I set all three parameters in a separate parametric file as a
part of the initial data collected. These three (Airport Elevation, Horizontal
Surface Elevation, and Conical Surface Elevation) are specific to the airport and
can be considered as global variables in the process down the road. Besides these
given parameters, the FAA provides detailed descriptions for every surface and
how they connect with each other. This last point is important to note because
every surface is dependent on one or more of the other adjacent surfaces, and,
consequently, provides a suggested order for development for each one: Primary,
then, Approach, Transitional, Horizontal, and finally, Conical.
Of course, this order is not mandated as being absolute. Rather, it is
strongly recommended as an order for systematic development as this is how the
imaginary surface order is followed from the centerline runway outside, and
described as the order of their dependability on each other.
As it was stated earlier, there are several types of runways that have
different design parameters for their imaginary surfaces. Consequently, imaginary
surfaces may have the same shape but different dimensions, and their
combinations will result in slight differences for different runways in the final
model. One of the main goals of my research was to discover a general approach
for building a Full PART77 model for a single runway, intending to apply the
same method to all remaining runway types. I began by investigating a number of
approaches and discovered in my research that there were a series of significant
points (SIGPOINTS) that actually describe all the imaginary surfaces of a
particular runway.
Further, I found that by identifying and connecting these points in a
specific order, it was possible to construct the boundaries of all of the imaginary
surfaces. Actually, the rules for imaginary surface creations were transformed into
simple mathematical equations. In addition, I sequenced the computing of these
equations, and in the end, I was able to describe the complete process solely in
terms of mathematical equations and logical rules.
The following summary is the sequence I implemented in order to develop
a single runway PART77 model from survey data:
1. Find and list the description of every surface, using the FAA
documentation for the specified runway.
2. Define the rule set in FAA PART77 documentation for these
imaginary surfaces.
3. List all the known parameters and given variables.
4. List all the missing parameters and variables needed to complete
imaginary surface development and calculations.

5. Design the mathematical equations needed to calculate the significant
6. Calculate the significant points.
7. Design the procedural logic to develop the imaginary surfaces for a
single runway.
8. Develop an automated procedure for implementation of equations and
logical procedures.
9. Develop a model and database for a single runway.
10. Develop a model and database for the whole airport this step was
added later after all existing runway PART77 models were developed.
The description and rules for constructing imaginary surfaces can be found
in FAA Federal Aviation Regulation 49 CFR PART77 (January 2007). For
example, the definition of a conical surface is:
A surface extending outward and upward from the periphery of the
Horizontal surface at a slope of 20 to 1 for a horizontal distance of
with its elevation represented as:
[15] CSE = HSEL + CSL *CSS
Where CSE is the conical surface elevation, HSEL is the horizontal surface
elevation, CSL is the conical surface length, and CSS is the conical surface slope.
The underlying logic for creating a conical surface was to buffer the
existing horizontal surface 4000 feet and then assigns the conical surface
elevation to the outer edge of the buffered zone. This logic was entered as a series
of commands in Model Builder as shown in Figures 23 and 24. Here, I have
illustrated two variations of this tool, each dependent on a different ArcGIS

According to FAA Federal Aviation Regulation 49 CFR PART77 (January
2007), the definition of a primary surface is:
Primary surface: A surface longitudinally centered on a runway. When
the runway has a specially prepared hard surface, the primary surface
extends 200 feet beyond each end of that runway; but when the runway
has no specially prepared hard surface, or planned hard surface, the
primary surface ends at each end of that runway. The elevation of any
point on the primary surface is the same as the elevation of the nearest
point on the runway centerline. The width of a primary surface is:
(1) 250feet for utility runways having only visual approaches.
(2) 500 feet for utility runways having no precision instrument
(3) For other than utility runways, the width is:

(i) 500 feet for visual runways having only visual approaches.
In keeping with the definition, a primary surface will always be
represented as rectangle around the runway. The width and the length are the
dimensions of the rectangle, as given, depending on the runway type. And its
surface can be described by its four comer points, in which its spatial coordinates
are computed from its runway end points. By default, the elevations of primary
surface points (PSP) are tied to the relevant runway end elevation. Four comer
PSP spatial coordinates are the function of the nearest runway end point (RWEP)
X and Y, the Az, and the distance from this point to PSP.
[16] X,Y(PSP) f (Xrwend, Yrwend, Az, Distance)
Where Xrwend, Yrwend are known, and Az and Distance from nearest runway
end point to primary points can be easily computed.
In Table 2 below are shown two variables, the primary surface distance
(PSD), and the primary surface azimuth (PSAz) computed for all the runway
types. Having all comer points for PS, we can connect them to delineate the
primary surface polygon. The built-in GIS tool, ASCII 3D to Feature Class, and
the default setting Polygon were used'to create this polygon.
Primary Surface Parameters
Given Calculated
Extent PSE Width PSW Distance PSD Azimuth PSAz
0 250 250 0
0 500 500 0
0 1000 1000 0
200 250 320.156 51.3402
200 500 538.516 68.1986
200 1000 1019.8 78.6901
Although the logical method of developing PS just described here is not
difficult to accomplish, it is important to note that the right parameters must be
used, that is, those based on the dependency and connection between
Extent/Width and runway types as presented in Figure 25.

Primary Surface Parameters

Using the layout illustrated in the preceding diagram, it is easy to
determine the PSE and PSW and, consequently, to calculate the PSD, PSAz, and
PSP. However, the four comer points are not sufficient to accurately describe the
correct Primary Surface, because as their elevations are based on runway end
elevations, the primary TIN will not follow the rule established by the FAA and
will not include the runway centerline points. The rule states:
The elevation of any point on the primary surface is the same as the
elevation of the nearest point on the runway centerline
Consequently, the other adjacent surfaces as Inner Transitional surface
will also be affected, and the overall result will yield the wrong width for the
Inner Transitional surface. It will either be too narrow or too wide, so in either
case, incorrect. Figure 26 illustrates both of the Primary surface TINs, based on
end points (figure a) and based on all points (figure b). I used the same color ramp
and the same classification method, setting the same break values to present both
TIN surfaces, to clearly illustrate the differences between the two.
a. Primary surface polygon and PS TIN
based on four PS points
Primary surface polygon, and PS TIN
based on four PS points and offset
centerline points

As a result of the process, the following questions surfaced: Why did this
happen? and Why are the TIN surfaces different?
With a closer examination of the break value statistics for each TIN
surface (see APPENDIX B Primary Surface), we can see that in figure (b) the
minimum elevation is 5315.959ft and the maximum elevation is 5325.749ft. This
differs markedly from the elevations in TIN surface based on runway end points
only, where the minimum elevation is 5321.07ft and the maximum elevation
5325.479ft. I began to investigate why a difference of almost 5ft existed between
the two minimum elevations.
The answer I found was that a vertical curve comes into play. In other
words, the PS polygon based on runway end points did not represent the correct
primary surface, resulting in an incorrect calculation of the TIN. This, of course,
has a significant impact on the adjacent surface located on each longitudinal side
of primary surface, as the adjacent surface is the Inner Transitional surface, which
typically tends to be the most critical surface for development near a runway. To
alleviate this problem and to make the Primary Surface a more accurate boundary,
I was required to describe this surface as series of points calculated from the
centerline points and offset, according to the primary surface width on each side
of centerline runway. Thus corrected, the TIN model of the Primary surface
provided the needed data for accurately delineating the internal boundary of Inner
Transitional surface.
Subsequently in the process, all of the imaginary surfaces were
constructed using this same type of logic. And as different runways had different
parameters, each type of imaginary surfaces kept the same shape while having
different parameters and elevations for their significant points. To view an
illustration of all surfaces, points and significant points in the project, see
Appendix C SIGPOINTS and Surfaces. Using this strategy, I was able to
develop the needed equations for each parameter, which were not given, but were
necessary to calculate the SIGPOINTS spatial position. Surface parameters and
their abbreviations are presented in detail in Appendix E Surface Parameters.
Parameters were then translated into variables for mathematical equations.
The variables given in red are those that have been derived from FAA
documents [1], [2], [31], [32],
The variables in black are missing data needed to complete the
SIGPOINTS calculation process.
The underlying principle is that every point from a SIGPOINT array can
be presented as Name, X, Y, Z. Where the Name or point number describes its
designation and X, Y, and Z are spatial coordinates of this particular point. The
calculation of spatial coordinates is tedious, but not unsolvable, as all surfaces are

based on the runway centerline, which is presented as series of points with their
X, Y, and Z coordinates. Therefore, the coordinates of significant points are a
function of the coordinates of existing runway end points (RWEP), centerline
points (CLP) and imaginary surface descriptions (ISD), as follows.
[17] SIGPOINTS (X, Y, ZJ = F(R WEP(x, y, z), CLP(x, y ,z),ISD)
The full PART77 model with SIGPOINTS is presented in Figure 27, with
each imaginary surface designated by a different color.
The Primary surface is represented in black, the Conical surface in yellow,
the Horizontal surface in green, the Approach surface (including both parts 50:1
and 40:1) in purple, and the Transitional surface (including Out and Inner
Transitional surfaces) in pink.
The figure clearly shows that all surface boundaries are described as lines
and arcs and each significant vertex is presented as a point. These points are the
SIGPOINTS. With defined rules and parameters, mathematical equations were
developed to calculate the missing variables. After the missing variables were
calculated, the equations to calculate the spatial values (PN, X, Y, Z, Code) of the
significant points were developed. Finally, to run all the calculations, an Arc
Object tool was developed that included all the equations. The entire process is
illustrated in Figure 28.

FINAL DATA TXT FILE U .........' **"*

The green rectangles represent given parameters and survey points, grey
rectangles represent calculation steps, and cyan rectangles present the sets of
SIGPOINTS that are combined into one single text file at the end. This text file
has all SIGPOINTS points listed in format PN, X, Y, Z, and Name. However, this
file was not the file used to create the surfaces. When these points were uploaded
into Arc Map, they still remained a list of points and needed to be connected to
delineate the imaginary surface boundaries. To do so, I described the final step of
the PART77 development process as a series of procedural logic based on the
ArcGIS built -in functions.
However, before presenting this final step, I would like to take the time to
identify the problems that I had faced, and which affected the correctness of the
final model. In Chapter 5 (Data Collections and Acquisition), I described a
problem that arose when determining the intersection line (IL) between Inner
Transitional Surface (ITS) and Horizontal surfaces (HS). As shown in Figure 9, a
difference of 476 ft existed above this line when comparing the results of both
development methods (from runway end points and from centerline points). If I
calculated from the centerline points, irrespective of the Primary surface that is
based on it, the points for this intersection line result in a more accurate
calculation. So, following the description for the inner portion of the Transitional
surface given in [1], [31], and [32], I determined that this line must have the same
elevation as the Horizontal surface, and further that the line be presented as series
of points, which have an elevation equal to HSEL. Therefore, the position of IL
points were as follows:
[18] (HP) Xp, Yp =f (CLP (X, Y,Z), HSEL)
Where, X and Y are the coordinates of centerline point, the Horizontal surface
elevation (HSEL), and the Transitional surface slope (TSS).
To expedite the calculation process of the spatial position of these points, I
developed another tool, writing mathematical functions as Visual Basic (VB)
code. The results could then be exported as a comma-delimited file, listing a
series of points (PN, Xp,Yp, Zp) for the left and right intersection line. I found
this way to be better than any other way I founding use in existing software
applications because it calculated all possible intersection points for each side
based on actual centerline point data. And, as previously pointed out, all existing
applications currently on the market use insufficiently simpler trigonometrical
methods to calculate both ends of this intersection line based on runway end
points and draw the imaginary surfaces. The method I developed also shows that
if the runway is not a straight line and has a consistent slope from end to end, the
subsequent Primary and Inner Transitional surface calculations would not be

correct. Hence, the distance between internal Inner Transitional surface and
Intersection Line would differ significantly, yielding incorrect data results.
Further, as the area bordered by ITS and IL is the most critical for new
development, it is also the area of greatest interest to planners. Thus, I cannot
underscore enough the importance of performing the calculation using the method
I just described. This method, the SIGPOINT computation process, is illustrated
in Figure 29. It shows the input data and resulting files based on SIGPOINTS and
ILS calculation methods. Several files were generated as a result of these
calculations, and all of these files were compatible with ArcGIS and could be
imported as Multipoint, PolylineZ, and PolygonZ feature classes.
To keep the process from becoming unnecessarily complicated and to
support the goal of clearly defining the steps in the process so as to streamline the
entire process, it is important to split the calculated data into several files. After
the SIGPOINTS were computed and the resulting files were generated, I
completed the development of full PART77 model as TIN and Raster surface

models, by using all the possible built-in functions of ArcGIS and developing a
workflow in Model Builder. The diagram describing the full process is presented
in Figure 30. Full and readable image of Figure 30 is presented in Appendix F -
Model builder flowchart to create one runway PART77 model from SIGPOINTS.
Finally, the process of developing a GIS database, the TIN and Raster
models for one runway PART77 was completed, automated and streamlined. The
entire process can be divided into two essential parts: calculating SIGPOINTS to
include ILS points, and using Model Builder to run the process and create the
model. The entire process is based on one single input data file that includes
runway centerline points. With the current version of my tool, all parameters can
be hard-coded, alleviating the potential for human error in the data entry process.
By using my method, any planner will be able to create a GIS PART77 model for
any single type of runway.

6.4 Full airport PART77 model
As a final element to my research, I developed another tool based on
Model Builder that will enable airport planners to combine all existing single
runway PART77 models and create a full airport PART77 model. An outline of
the process is illustrated in Figure 31.
The results include TIN and Raster models for the whole airport. Using
this model, planners can conduct analyses based on the FAA AC for model
zoning ordinance [5] and other necessary calculations. Further, this model covers
a very large area and consists of the maximum restrictive PART77 surfaces. This
model also can be transformed into the map required by the FAA as part of the
Airport Master Plan. Before this model was proposed, map creation typically
required a team of consultants, which added extensively to the cost and time-
commitment for a project. With the newly proposed methods, any airport planner
or GIS analyst can create the map within a few hours. It saves both time and
money, and more importantly, if something appears incorrect, it can readily be

fixed without increasing the scope of a consulting project or its costs.
6.5 Verification and sensitivity of the model
I would like to address the verification and sensitivity of the model in
relation to potential errors in the model itself. In other words, How can the
process validate the correctness of the modeP.
First, the model employs mathematical functions in its calculations, where
two plus two indisputably equals four. Second, the FAA advisory circular clearly
describes the means for calculating the most important values for imaginary
surfaces. This means serves as the underlying source of data for the mathematical
equations. Third, when modeling the shapes of imaginary surfaces, no variances
exist. The shapes dimensions may differ from runway to runway, but the shape is
always preserved. To provide evidence for the models validity, lets take another
example dealing with Primary surfaces.
In this example, we will examine a Utility runway in two different
contexts: Precision-Instrument Approach and Non-precision approach. In both
cases, the spatial coordinates of runway endpoints that need to be computed, the
four PSP SIGPOINTS, are identical. So, we will use the same data for two
different variants and assume that runway type is different as described in the
following tables and illustrated in Figure 32.
1. RW CAT III, Utility, Precision-Instrument Approach, hard surface
From Table 1:
Primary surface extend PSE = 200ft
Primary surface width PSW = 1000ft
From Table 2
Distance to PSP from nearest RWEND = 1019.8ft
Azimuth to PSP from nearest RWEND = 78.6906 Decimal Degrees
Coordinates for four PSP:
2. RW CAT III, Utility, PIA, Non-precision approach, hard surface