Hobbs_RMMESthesis2016.pdf

Media

Part of Bird & Wildlife Patterns at SEATAC Airport Utilizing a Foreign Object Debris Detection System

extracted text

BIRD & WILDLIFE PATTERNS AT SEATAC AIRPORT UTILIZING A FOREIGN
OBJECT DEBRIS DETECTION SYSTEM

by
Ryan M. Hobbs

A Thesis
Submitted in partial fulfillment
of the requirements for the degree
Master of Environmental Studies
The Evergreen State College
December 2016

©2016 by Ryan Hobbs. All rights reserved.

This Thesis for the Master of Environmental Studies Degree
by
Ryan Hobbs

has been approved for
The Evergreen State College
by

________________________
Kevin J. Francis, Ph. D.
Member of the Faculty

________________________
Date

ABSTRACT
Bird & Wildlife Patterns at SeaTac Airport Utilizing a FOD Detection System
Ryan Hobbs
Since 1990, 143,000 reported wildlife-civil aircraft strikes have occurred in the United
States. These strikes have resulted in 26 human fatalities, 174 injuries, 66 destroyed
aircraft, and billions of dollars in damages. Birds and wildlife account for nearly 40% of
all foreign object debris (FOD) incidents and pose a risk to both themselves and humans.
New radar technologies to detect FOD are rapidly developing to replace human
observation as the primary detection method. As part of the wildlife management
program at Seattle-Tacoma International Airport (SEA), a FOD radar system was
installed to generate a better understanding of bird and wildlife patterns and to mitigate
strikes. To determine the capabilities of FOD radar, data analyses from a recently
installed system SEA found large numbers of birds (n=668) and wildlife (n=4,072),
which includes insects, along the center runway over a nine-month period. The current
years strike database also indicates a low number of strikes (n=52) compared with the
average yearly count (n=101) from 2005-2016 and the lowest in a decade. This low
number suggests FOD radar successfully mitigated strikes at SEA and the collected data
is useful in actively guiding future management plans.

Table of Contents
Introduction……………………………………………………………………………..…1
Seattle-Tacoma International Airport…………………………………………...2-3
Wildlife Foreign Object Debris…………………………………………………3-4
Xsights FODetect……………………………………………………………….4-5
A Review of the Literature………………………………………………………….…….6
Introduction……………………………………………………………………..6-7
Habitat and Wildlife Management at Airports………………………………...7-10
Wildlife Threats to Aircraft…………………………………………………..11-13
Economic Impacts of Strikes to Aviation Industry………………………………13
Avian Monitoring with Radar at Airports…………………………………....13-16
Methods……………………………………………………………………………….…17
Setting and Study Area…………………………………………………………..17
Methodology………………………………………………………………….17-18
Radar Data…………………………………………………………………....18-19
Results……………………………………………………………………………………19
Descriptive Data Analysis……………………………………………………19-29
Discussion & Further Research………………………………………………………….30
Technical and Data Management Recommendations………………………..30-31
Further Research……………………………………………………………...32-33
References……………………………………………………………………………33-36
Appendices…………………………………………………………………………...37-40

iv

List of Figures
Figure 1. Monthly wildlife detections
Figure 2. Estimated wildlife sizing chart
Figure 3. Map displaying wildlife sizing index along the center runway
Figure 4. Monthly bird detections
Figure 5. Estimated bird sizing chart
Figure 6. Map displaying wildlife sizing index along the center runway
Figure 7. Heat map displaying wildlife population densities along center runway
Figure 8. Heat map displaying bird population densities along center runway
Figure 9. Yearly reported strikes between bird/wildlife and aircraft

v

Acknowledgements
My humble thanks to all those that supported me during the course of completing
this work. Thank you to Kevin Francis, MES Director, for his constant feedback and
guidance during the writing process. Thank you to Steve Osmek and the team at the Port
of Seattle’s Seattle-Tacoma International Airport for working with me and allowing me
to use their data as a case study. Thank you to my cohort and the students that supported
me along the way, especially Hannah Trageser, who opened up the opportunity to work
with the Port of Seattle. Thank you to Mike Ruth for opening up the world of GIS to me
and my colleagues at The National Audubon Society for their creative ideas and support.
Thank you to my family who has supported my constant need to further myself and my
education, without their support, food, and housing, I would not have survived this
process. Thank you to Haley Danger Heck for her tremendous support in both the writing
process and overcoming the emotional strain I experienced throughout the course of this
project. And lastly, a major thank you to my cat Lucien, for without his unconditional
support and pestering, I may have finished this thesis much, much earlier.

vi

Introduction
Wildlife and airports have a complicated relationship. Airports are appealing to
birds and wildlife for a variety of reasons. they are often located in open, flat areas, close
to major rivers or water sources, providing habitat for many species. Migration paths can
cross airports and be seen as an appropriate location for a feeding stop. Airports may also
lie adjacent to important bird and mammal habitat, causing crossover as species move
around airports.
In contrast, the threat of birds and wildlife mixing with aircraft has proven to be a
recipe for disaster. In 2009, an incident involving US Airways Flight 1549, which
collided with a flock of geese shortly after takeoff, resulted in loss of thrust in both
engines. The pilot safely landed the aircraft in the Hudson River, preserving the life of
the 150 passengers1. While this was not the first account of a major strike incident
between birds and aircraft, it was widely reported and garnered nationwide attention. This
has led to a wider awareness of the need for mitigation strategies to limit bird and wildlife
interactions with aircraft.
Many steps have since been taken to minimize the impact of birds and wildlife on
the aviation industry: wildlife management practices such as hazing birds, restricting
entrance through fencing, and changing the physical landscape of airports through
modifying grasses, covering water sources, and proper grading to reduce runoff pools.
Airports have also implemented technology to provide constant monitoring of bird and
wildlife movements. As the field of airport wildlife management continues to grow, The
1

US Airways Flight 1549 was a bird strike incident that received significant media attention. Since the
event, multiple books and feature films detailing the incident have been released. The complete details of
the strike and crash can be found at the National Transportation Safety Board website
http://www.ntsb.gov/investigations/AccidentReports/Pages/AAR1003.aspx

1

Port of Seattle’s Seattle-Tacoma International Airport (SEA) is taking strides to remain
ahead of the curve. In late 2015, SEA installed a radar system to detect and deter birds
and wildlife on their center runway.
This thesis provides an analysis of the bird and wildlife distribution along the
center runway, as well as an assessment of the ability of the radar system to mitigate
strikes between aircraft and birds. This analysis will be conducted with data collected and
documented by the Xsight FODetect radar system and Airport Operations (AO) and the
(FAA) strike database. Preliminary results suggest that the radar system will reduce
strikes

Seattle-Tacoma International Airport
SEA became the first airport to hire a full-time wildlife biologist in 1970s with the
aim of developing an ecological approach to airport wildlife management. With its wideopen spaces and buffer areas, SEA is an appealing habitat to wildlife. Identifying and
eliminating specific wildlife attractants is key to wildlife management. The airport
currently employs a comprehensive wildlife management program to deter wildlife from
congregating on site. In terms of runway safety, one of the most recent wildlife
management tools employed by SEA is installation and use of a foreign object debris
(FOD) radar system along the center runway (Appendix 1). These radar systems were
installed to identify both wildlife and non-wildlife threats on and along the runway. On
January 1, 2016, SEA officially became the second U.S. airport to utilize FODetect radar
systems; the installation of FODetect coincides with the center runway project and seeks
to increase foreign object debris (FOD) awareness.

2

The need for a constant runway monitoring stems from the high volume seen at
the airport. SEA runs 855 daily operations with over 42 million passengers arriving and
departing in 2015, an increase of 12.9 percent over 2014. These large numbers have
resulted in SEA being ranked the 13th busiest U.S. airport by passengers in 2014.
Operationally, SEA was ranked as the 22nd busiest U.S. commercial service airport in
2013 (Port of Seattle, 2015). Frequent air traffic also requires that SEA maintain the
highest level of operations efficiency and safety. In order to increase aviation safety,
developing an understanding SEA’s attraction to wildlife is important.
This understanding of the attractiveness is one that has long been underway for
the wildlife team at SEA, as they have developed an extensive management plan (Port of
Seattle, 2004). This plan lays out ways for the airport and the wildlife team to manage
and modify the attractants on Port of Seattle (POS)-owned land as well as their strategies
for working with non-POS land owners in the area. The management plan also involves
an in-depth description of the airport ecosystem and coinciding management practices:
airfield structures, airport building projects, abandoned structures, non-airport land-use
projects, water management, wetlands, lakes, stormwater and temporary pools, vegetation
management, streamside vegetation, grass types and management, ornamental
landscaping, and food/prey-base management. The document also includes a priority
timeline for management projects. A tool of this sort would be beneficial to all airports as
it describes the bird and wildlife attractants and strategies for reducing the airport’s
appeal.

3

Wildlife Foreign Object Debris
Foreign object debris (FOD) can be defined as any object in an inappropriate
location that may cause damage or injury to aircraft, equipment, and airport personnel.
Items often identified as FOD include rocks, loose hardware, chunks of pavement, cargo,
sand, and wildlife. These items may cause damage from direct contact with aircraft,
including window and body collisions or from intake into the jet engines, resulting in a
variety of damage. Between 2000 and 2013, wildlife FOD resulted in 66 destroyed civil
aircraft, 26 human deaths, 174 injuries to passengers and grounds crew, thousands of
dead wildlife, billions of dollars in damages, and millions of hours in aircraft downtime.
With this threat constantly lingering at SEA, the wildlife staff has been active in limiting
bird and wildlife presence at the airport and FODetect is the most recent line of defense
in mitigating bird and wildlife strikes with aircraft. They believe this will help them
further develop methods for understanding the patterns of birds and wildlife along the
center runway.

Xsight’s FODetect
FODetect hybrid radar systems use Surface Detection Units (SDU) that utilize
both a radar and camera unit to identify objects on and near runways. In general terms,
radar works by sending out pulses of radio waves, which bounce off an object and return
to the point of origin. Based on how long it takes for the signal to come back from the
object, the user can determine the shape of the object and distance from the source. This
system functions at 77 GHz, which is considered highly accurate and uses a greater
bandwidth or increased object detectability and resolution. The radar unit sends a

4

constant stream of information, with the ability to conduct a full 360° scan of the runway
and the field areas alongside, while identifying objects and alerting staff of potential
threats. As part of the Federal Aviation Administration’s (FAA) regulations on FOD
monitoring, the airport staff conduct three daily FOD checks (two more than the FAA
requires). This system adds another layer of protection and efficiency with its constant
scan.
A key feature of FODetect is the ability to detect birds and other wildlife. It
utilizes Xsight Systems BirdWize and alerts Aviation Operations (AO) of potential
threats. To increase accuracy, the system uses both a radar algorithm and image
processing software to detect FOD at close-range. When an object is detected, an alarm
triggers AO crews who use the unit’s camera to determine threat level and whether or not
further action is required (Appendix 2). If further action is required, the AO crew takes
the necessary steps to ensure safe paths for aircraft. This may involve rerouting landings
to different runways or sending out crewmembers to handle FOD cleanup. The system
also includes a sound generator to startle birds and other wildlife.

5

A Review of the Literature
Introduction
Literature regarding wildlife at airports has broadly focused on management and
threats to aircraft. Current findings in this research describe best practices in order to
reduce wildlife populations at and around airports. While much has been explored in
these areas, literature on monitoring and mitigating wildlife FOD leaves much to be
desired. Since the use of radar to detect FOD on runways is a fairly recent trend, the
established research focuses on the detection range and proposals for improved
equipment, rather than the effectiveness of the system to deter strikes. Current and
increased emphasis on FOD detection has led to the FAA creating requirements for FOD
technology systems in 2009.
The two most significant factors driving the increased interest of FOD detection
technologies have been the safety threats to humans, birds, and wildlife and the economic
impact to the aviation industry. Bird and wildlife FOD have been present on the radar
since the 2009 incident with US Airways Flight 1549 and the new technologies being
rolled out seek to minimize the significant events involving bird and wildlife interactions
with aircraft. Reducing these events will not only increase aviation safety, it will also
reduce the major economic impacts these strikes present.
The technology currently employed at SEA is a hybrid system combining radar
and optical features. It is automated but allows for complete control by an operator and
serves as a constant set of eyes on the runway. While the FAA produced a study
assessing FODetect’s ability to detect FOD on runways, the study neglected the system’s
6

ability to detect and mitigate potential wildlife FOD in the field areas near runways.
Much is desired in the understanding of how FOD radar will serve its purpose at airports
and to what degree it is capable of performing. This study will provide the data and
analysis needed to fill in the gap left by the current bird and wildlife FOD studies.
This literature review summarizes and examines the available literature as it
pertains to wildlife management and wildlife threats to aviation safety and how the
development of technologies assists in mitigating wildlife strikes. While current
management trends provide means of limiting wildlife at airports, reducing wildlife
access to airports required constant adaptations and modifications. The issue of wildlifeaircraft strikes with an emphasis on birds will also be explored. Economic impacts to the
aviation industry will be examined, as will current wildlife radar uses, and the abilities of
FOD detection systems. It is also important to address the viability of human observation
of FOD near runways to illustrate the potential for FOD detection technology to increase
awareness of wildlife near runways.
Habitat and Wildlife Management at Airports
The Port of Seattle owns SEA property and the Port of Seattle’s Century Agenda
(2012) was to become the greenest, cleanest, and most energy efficient port in the nation.
With its wide open spaces and buffer areas, SEA is an appealing habitat for wildlife.
Identifying and eliminating specific wildlife attractants is key to wildlife management.
The airport currently employs a comprehensive wildlife management program to deter
wildlife from congregating on site. According to the Port of Seattle’s wildlife
management plan (2004), guided by several federal regulations, a variety of management

7

practices are employed. Coyotes, dogs, deer, and other large mammals are physically
excluded by fencing. However, this has minimal effect on small mammal and avian
populations. Some daily airport activity, such as driving on the infields between runways,
has resulted in temporary pools during times of precipitation, leading to increased duck
and shorebird populations. SEA currently plans for these areas to be graded for water
runoff (SEA, 2004). Nearby attractants are covered with netting to eliminate waterfowl
congregations.
In order to reduce wildlife strikes, understanding the complexities of wildlife and
the attractants luring them to airports is vital. One of the first steps in managing wildlife
is recognizing wildlife movements, which vary daily, seasonally, and annually. Factors
that contribute to variation range from broad biological and ecological characteristics
such as migration, foraging, and habitat preferences (Belant, Washburn, & DeVault,
2013). Belant et al. (2013) also suggest five categories of animal movements including
foraging, dispersal, migration, movement to rest locations, and territorial defense.
Movements may also be dictated by resource abundance (Fretwell & Lucas, 1970) and/or
niche filling (Krebs, 2001). Species will often follow the movement of prey or spread out
to fill a niche in ecosystems, which may allow for successful population movements.
Lower altitude strikes with birds may result as aircraft are in the direct flight path
of species known to quickly fly through lower elevations of their habitat (Klem, 1990,
Klem 2010, Bayne et al., 2012). One contributing factor to bird strikes in this area
involves spooked flight, which occurs when birds are startled by aircraft or loud noises.
Along with spooked flight, play flight—in which birds chase one another, resulting in
erratic, frantic flight paths—is likely to contribute to strikes. In more recent years,
8

commercial aircraft have turned to quieter two-engine aircraft, moving away from three
to four-engine aircraft. Two-engine aircraft, while quieter and more reliable, are also
more vulnerable to the issues brought about by multiple ingestion events, much like the
US Airways Flight 1549 incident.
These quieter engines may result in bird and wildlife’s lack of awareness of
oncoming aircraft (FAA, 2013). Past experiments involving engine recording playbacks
have shown birds and wildlife would need a greater escape distance from the newer,
quieter, and faster engines than the previously used and noisier engines (Solman, 1981).
While some species may not be affected by aircraft noise, others may not be able to
distinguish incoming aircraft from background noise at airports (Conomy et al. 1998).
As part of SEA’s management plan (2004), the grass around runways and taxiway
markers must be kept at a 3-inch maximum height. An abundance of grasshoppers is
often found in the area, which are seen as a wildlife attractant, particularly to crows
(Corvus brachyrhynchos) and raptors (SEA, 2004). Reducing the height of grass by
mowing should serve to reduce grasshopper populations by limiting their preferred
habitat of taller grass. Airport biologists believe mowing itself to be an attractant as it
opens up food sources, revealing available food sources. It has been reported (Buckley
and McCarthy, 1994) that laughing gulls (Larus atricilla) at John F. Kennedy
International Airport in New York only fed on Oriental Beatles (Anomala orientalis) in
areas with shortgrass, even though the abundance was equal in tallgrass areas, suggesting
that food availability rather than food abundance is key to habitat use. Also supporting
availability over abundance is the use of shortgrass areas by red-tailed hawks (Buteo
jamaicensis) and rough-legged hawks (B. lagopus) to prey on meadow voles (Microtus
9

pennyslvanicus) even though population densities in shortgrass are far less rich (Baker &
Brooks, 1981). Thus, mowing is prescribed at times when birds are most inactive.
Monitoring wildlife attractants such as insects, trash, and debris is a constant
activity at SEA. These types of attractants often bring in crows, European starlings
(Sturnus vulgaris), raptors and waterfowl. SEA also has a well-established raptor
monitoring and relocation program (Raptor Strike Avoidance Program) to remove
potential threats from the airfield (Anderson & Osmek, 2005). Pesticides are used in
instances where other methods for controlling insect, weed, small mammal, and some
bird populations are exhausted. Grass seeds that are known to be unattractive to many
avian species are also used as a means of reducing bird populations (SEA, 2004).
The above mentioned practices of SEA have also been recommended by many
experts (Blackwell, DeVault, Fernandez-Juricic, & Dolbeer, 2009; Cleary & Dolbeer,
2005; DeVault, Belant, Blackwell, & Seamans, 2011; Cleary, Dolbeer, & Wright, 2006),
though a single recipe for management across all airports is unrealistic. Emphasis on
habitat management, reduction of attractants, implementation of deterrents, and the
hazing and harassing of bird species are common management recommendations. Along
with these practices, it is also recommended that survey data of species be used to
identify management priorities and thus reduce hazards to aviation safety (Blackwell,
Schmidt, & Martin, 2013). Sufficient survey data can be used in the development of
runway protection zones, which allow for targeted methods of wildlife management for
particular species.

10

Wildlife Threats to Aircraft
Wildlife, particularly birds, have been a threat to aircraft since humans first took
to the sky, with the first strike being recorded by Orville Wright in 1908 and the first
death from a strike occurring four years later (Thorpe, 2003). While strikes with wildlife
do not always result in damage to aircraft or human injuries, predictions suggest
increased frequencies in airplane/wildlife strikes (Coccon, Zucchetta, Bossi, Borrotti,
Torricelli, & Franzoi, 2015). The FAA has also predicted an average increase of
passenger growth by 2.1% per year over the next 20 years though the fleet growth is
minimal at just 0.02% a year. However, with this minimal increase in fleet growth,
aviation hours are projected at an average increase rate of 1.2% per year (FAA Aerospace
Forecast, 2016).
A massive increase of reported bird and wildlife strikes can be seen in the FAA’s
strike database. In 1990, 1,851 wildlife strikes were reported, while 2013 had 11,315
reported strikes. This can be attributed in part to a general rise in reporting, however
researchers have found that roughly 25% of all wildlife strikes are reported and attribute
low reporting percentages to the pilot or aviation operators not knowing a strike occurred
or to avoid logistical inconveniences (Cleary at al., 2006; Dolbeer & Wright, 2009;
Linnell, Conover, & Ohashi, 1996, 1999).
The physical characteristics of airports can also contribute to strike occurrence. It
has been found that wildlife threats are common at airports with expansive grasslands
(Dolbeer, 2000) and SEA fits this criterion. Beyond the physical characteristics
contributing to strikes, Cleary et al. (2006) found that increased populations, in part from

11

conservation efforts, has magnified wildlife threats to aircraft. Dolbeer and Eschenfelder
(2003) also highlighted rising populations of species that pose a higher level of threat due
to their body size (greater than 8 pounds). Using physics, one can determine the amount
of energy a bird of this weight would transfer to the airplane by squaring the speed
difference. As the weight increases, so does the energy. Measuring force is more dynamic
as the amount of time the bird is in contact with the aircraft is considered in this equation.
Not all birds and wildlife are considered a major threat to an aircraft and there is a
broad range of threat levels posed by individual species. When assessing the risks
associated with wildlife species, Dolbeer et al. (2000) used 21 species or species groups
that had 17 or more strikes from 1993-1998. Three variables (percent of strikes causing
damage, major damage, and effect-on-flight) were determined to assess wildlife threat
risks. The species listed as most hazardous to aircraft were artiodactyls (mostly deer) and
large birds such as vultures (Cathartidae), geese (primarily Branta canadensis), hawks,
and eagles. The species with the least risk were typically smaller avian species such as
sparrows (Emberizidae), swallows (Hirundidae), and blackbirds-starlings (IcterinaeSturnus vulgaris). It is important to note that deer and vultures, both posing high risks
were also lower in strike percentage than many of the other species. While those on the
lower end of the spectrum were smaller in size, they accounted for a larger strike
percentage, illustrating their hazard to aircraft.
Though strikes rarely result in injury or death to humans, there have been
instances of human fatalities. In 1962, a flight departing from Boston’s Logan Airport
struck a flock of European starlings, resulting in 62 human fatalities (USDA, 2012). In
Ethiopia, 35 human fatalities were attributed to flocks of speckled pigeons (Columba
12

guinea) and starlings in 1988 (Dolbeer et al., 2000). Globally, the FAA reports that since
1988, over 255 human fatalities have occurred and over 243 aircraft have been destroyed
from wildlife strikes. SEA’s wildlife biologist Steve Osmek stated that over 40% off all
FOD damage is caused by wildlife (S. Oskmek, personal communication, May 2015).
An assessment of bird strikes on aircraft found that 66% of damaging strikes
occurred at or below 500 feet. This range also was more likely to see strikes with raptors,
gulls, passerines, pigeons and doves (Dolbeer, 2006). In the report, it was also stated that
the months from July to October were the most common with strikes at or below 500 feet
in altitude.
Economic Impacts of Strikes to Aviation Industry
The estimated economic impacts from wildlife strikes vary and have been
estimated between $500 million (Cleary et al., 2004) to over $718 million per year by the
FAA (USDA, 2012). Civil aircraft downtime hours associated with reported strikes from
1990-2013 ranged from 55.6 to 388.4 and showed a mean of 119.4 hours. Repair costs
and other associated costs averaged at $167,100 and $29,852 respectively per report. In
total, there were 13,497 reports that listed damage (Dolbeer, Wright, Weller, & Begier,
2014).
Avian Monitoring with Radar at Airports
A variety of radar sensors are used to monitor birds: tracking radar, weather
surveillance radar, terminal Doppler weather radar, and high-resolution marine
surveillance radars have been used in ornithology to gather detailed flight patterns
including migration and foraging as well as population densities (Gauthreaux & Schmidt,
13

2013). This information can be used to guide specific management practices in
anticipation of and to mitigate specific behaviors. SEA currently employs three avian
radar (Sicom-Accipiter®) systems to detect potential wildlife hazards with a constant
focus on attractant areas. The radar can also scan upwards and outwards to a degree that
human observation cannot provide. These radar systems were installed prior to FODetect,
which has just begun scanning the center runway.
In 2004, the FAA began a program evaluating FOD detection systems (FAA,
2012). During 2007, the University of Illinois Center of Excellence for Airport
Technology (CEAT), which is part of the FAA’s Airport Technology and Development
Program, developed a performance assessment and the sensors were installed at Boston
Logan International Airport (BOS) in 2008. The system, FODetect, is developed by
XSight Systems, Ltd and was recently installed at SEA. The study assessed four major
categories and several subcategories, the first being basic functions. Under the umbrella
of detection performance was object detection, location accuracy, inspection frequency,
detection response time, surveillance area, and performance in weather. System
performance addressed alerts and alarms while system output covered detection data, data
presentation, and data management. Each of these categories are connected with
performance standards that must be met to be considered for use on the runway.
Selected targets evaluated consistency of the radar and optical target. Ten sensors
were installed, five on each side of the runway and they performed a paired scan of the
150-ft wide runway (Hericks, Woodworth, & Patterson, 2012). Performance tests
conducted in daylight hours on dry pavement in 2011 showed 100% detection rates and
FODetect was able to identify two small items roughly 6 inches apart. Tests on the sensor
14

provided results of initial object detection of 94% with an increase to 98% detection after
use of the camera. Average detection time was 60 seconds but were as low as roughly 30
seconds in some instances. Data alerts were found to be manageable with an Xsight
developed toolbox. The capabilities included location of FOD object, images of detected
FOD, image and video playback, current status (detected, confirmed, and retrieved), and
description of retrieved FOD (size, color, and category). The performance assessment
found that FODetect met the FAA’s “Airport Safety Self-Inspection” Advisory Circular
requirements.
Avian surveying and rigorous data management can form the basis for appropriate
management strategies and reduce avian threats to aviation safety (Blackwell et al.,
2013). Using scientifically sound data collection methods may also be necessary to
legally justify management actions, especially ones involving lethal control or in events
that involve human health and safety (Reiter, Brunson, & Schmidt, 1999; Conover,
2002). Improving the validity of data involves increasing precision and reducing bias.
Data that has been collected using tools to allow for precise measurements and accurate
accounts, are often considered trustworthy and reliable. Biases are systematic errors that
can influence the inaccuracies of survey data. Such biases may include observer ability,
behaviors of observer (serving as an attractant or repellent), weather, time of day, and
season (Blackwell et al., 2013). While Cleary and Dolbeer (2005) have outlined the
guidelines for avian surveys at airports, Blackwell et al., (2013) have also offered
strategies to improve survey accuracy. These strategies involve defining both an objective
and target species but also include modeling and strike risk as important factors in

15

analyzing survey data. However, variability and biases need to be taken into account to
reduce imprecisions.
The current literature regarding FOD radar is slim and often centered on the main
function of the radar itself. While these issues are important and justify the use of FOD
radar, it is necessary to examine the use of FOD radar and how it may contribute to
aviation safety as it relates to birds and wildlife. This thesis will address the lack of FOD
radar’s use in monitoring bird and wildlife along the runway by contributing a study of
the impacts FOD radar has on strike mitigation. It will also provide suggestions for ways
to use the data collected by the system to develop wildlife management plans as well as
how to improve on data collection.

16

Methods
Setting and Study Area
The Port of Seattle’s Seattle-Tacoma International Airport (SEA) lies in a
particularly dense urban environment, situated near Puget Sound and located 12 miles
south of downtown Seattle. With its wide open spaces and buffer areas, SEA is an
appealing habitat to wildlife and identifying and eliminating specific wildlife attractants
is key to wildlife management. The airport currently employs a comprehensive wildlife
management program to deter wildlife from congregating on site. In terms of runway
safety, one of the most recent wildlife management methods SEA is currently
undertaking involves the installation of a foreign object debris (FOD) radar system along
the entirety of the 9,426 feet long center runway. Each of the 101 radar/camera hybrid
units is placed along the edge of the runway with a vertical spread of 200 feet between
each unit. Spanning the width of the runway is roughly 137 feet between units. These
radar systems were installed in November 2015, with the aim of identifying both wildlife
and non-wildlife threats on and along the runway. The radar system officially began
operations on January 1, 2016 with planned downtime in June and late September
through October, 2016, during times of construction.
Methodology
The initial exploration of the data intends to identify the locations, frequencies,
temporal change, and size ratings of the wildlife and birds along the center runway. Excel
PivotTable analysis was used to answer the questions of monthly events, frequencies, and
size range of the two categories. ArcGIS software was used to identify bird and wildlife
17

hotspots and to differentiate the bird events by size, in order to identify individuals that
are within a high-risk range and their locations on the runway.
Radar Data
A constant stream of data is made available by the automated radar system. The
hybrid system uses a camera, which can be controlled by an operator, to further identify
objects on the runway. In regard to objects identified by the radar off the runway, no
images are automatically provided and it is required that an operator identify objects by
taking control of the camera and shifting its focus. Imagery taken by the camera also
displays a red rectangle to show the exact parameters that have been identified by the
radar (Appendix 3). The lead wildlife biologist provided the data by exporting an Excel
spreadsheet file from within the Xsight Systems FODetect software suite. The file was
then converted to a comma delimited (CSV) text file, which allows data to be easily
visualized and analyzed in ArcGIS software. The CSV contained a comprehensive list of
criteria as identified by the radar. For the purpose of this study, only objects identified as
birds or wildlife by the system or human operators were assessed.
Each sighting of an item by the radar is listed as a new event in the system. For
consistency purposes, each item will be referred to as an event. The data used in this
study spanned from January 1, 2016 through the morning of September 26, 2016. The
month of June accounts for only four days of observation as the runway was periodically
closed during this period. When determining the bird and wildlife hotspots, heat maps
were generated within ArcGIS Online (AGOL) to visualize data and identify areas as
high frequency event locations.

18

Results
Descriptive Data Analysis
The number of wildlife events between January 1 and September 26 was 4,072.
On a monthly basis, the wildlife counts, including insects, remained relatively low
through the winter and spring but spiked during the summer months (Figure 1).

Figure 1. Wildlife detections per month remained low through the first half of the year but quickly
spiked in July (*the system was disabled for most of the month of June).

The summer months averaged roughly 7.5 times as many events (1,108) compared to the
average of winter and spring (145 excluding June). Using the system’s archived images
and videos, September showed an influx of Mantises (Appendix 4), which may explain
the relatively low number of grasshoppers the radar and surveyors have reported this year
(mantises are a natural predator of grasshoppers).

19

In terms of size, the average wildlife event was 1.32 in2 (Figure 2), indicating that
the majority of wildlife were likely insects (Appendix 5).

Estimated Wildlife Size
4000

3772

3500

Frequency

3000
2500
2000
1500
1000
500

248
21

12

3

10.01-30

30.01-90

90.01-270

0
-1-2

2.01-10

Area (inches2)

Figure 2. The size of wildlife events was mostly low and was largely made up of soft bodied
insects. Some small mammals were identified and some clustering of wildlife as well as software
issues accounted for outliers. Negative values can be attributed to detection box errors.

A map depicting size distribution along the runway showed the smaller size range
distributed across the entirety of the runway with randomly scattered larger events

20

(Figure 3).

Figure 1. Map showing size distribution along the center runway, with the largest events
occurring near the center of the runway.

Rodents and other small mammals would see a size frame of roughly 5 to 20 inches. As
Belant, Washburn, & DeVault (2013) point out, understanding these changes through
21

time will allow for better management practices and the mitigation of bird and wildlife
strikes with aircraft.
Bird sightings, as determined by the system, totaled 668 events (Figure 4). In
terms of monthly variance, February (129) and July (123) involved the most events with
a disparity of events between the remainder of the months.

Bird Detections per Month
Feb
129

Jul
123

Sep
86

Mar
78

Aug
69
May[SERIES NAME]*
56
[VALUE]

Jan
41

Apr
30

Figure 4. Bird detections per month were highest in February and July with moderate sightings
throughout the remaining months (the system was disabled for most of the month of June).

Events sharply dropped from these high figures for the months of January (41), March
(78), April (30), May (56), June (56), August (69), and September (86).

22

Regarding object size, the average bird event was 14.09 in2 (Figure 5).
While this does not indicate the actual size of the bird, it does indicate the area of the
rectangle around the identified object.

Estimated Bird Size
350
292

300

Frequency

250
200
162
150
100
100

90

50
10
0
-1-2

2.01-10

10.01-30

Area

30.01-90

90.01-410

(inches2)

Figure 5. Bird sizes maintained low to medium values across the board, with some outliers
caused by multiple individuals contained within one event. The values were binned by size ranges
and frequencies are displayed atop each size range bar. Negative values can be attributed to
detection box errors.

A larger event points to either a large specimen or a group of birds all within the
identified rectangle, still serving as a relevant indicator of threat. The strike index
outlined by Dolbeer et al. (2000) can used in conjunction with the radar’s data in order to
rank threats along the runway. The size distribution map shows the majority of large

23

sightings in the north and central section of the runway (Figure 6).

Figure 6. Map showing bird size distribution along the center runway, with the majority of larger
events occurring near the center and northern portion of the runway.

In some instances, an event was identified though the rectangle failed to form and
remained as a straight line. Another issue with the radar’s detection ability came about
24

when only the feet of a bird was identified, causing the system to create a new, smaller
event, greatly skewing the size reading.
A key finding in the archival photos of the birds was a group of horned larks.
While it is not certain if the sightings were the subspecies, streaked horned larks
(Eremophila alpestris strigata), the wildlife team at SEA have contacted Washington’s
Department of Fish and Wildlife to make notice of the birds. Since streaked horned larks
are listed as threatened and a ground dwelling species that prefers wide open species with
little to no vegetation, the airport could be seen by the birds as a prime habitat. It was the
first such sighting and the discovery could indicate the need for major management
changes at SEA in the future. The use of archived photos supports Blackwell, Schmidt, &
Martin’s (2013) recommendation for using observational data to guide current and future
avian and wildlife management.

25

Wildlife incidences were greatest along the southern portion of the of the runway
though there was a relatively even distribution along the entirety of the runway (Figure
7).
Fig
ure
7. A
heat
map
disp
layi
ng
the
4,07
2
wild
life
even
ts
alon
g
the
cent
er
run
way
.
The
high
est
con
cent
rati
on
of
wild
life
occ
urre
d
alon
g
the
bott
om
port
ion
of the runway.

26

When the runway was divided into three vertical segments of 3,142 feet each, the
distribution showed 1,239 on the north end, 1,307 in the center region, and 1,526 on the
southern end.

Analysis of the bird data in AGOL (Figure 8) found the highest concentrations on
the north end of the runway (303 events), followed by a similar cluster near the center of
the runway (209 events), accounting for 77% of the total events between these two
locations. In terms of sizing, the larger events were also found in these hotspot regions,
with the middle portion of the runway containing similar values of medium to high threat
events while the southern extent of the runway contained low to medium threat events.
The SEA strike database indicated a decrease of strikes since the FOD radar
system came into use. Comparing 2016 to with previous years (Figure 8), the number of
strikes is considerately lower. It was also the lowest year since 2006, marking a dramatic
drop in strikes from 138 in 2015 to 52 as of October 2016. The average number of strikes
for the past 12 years (including 2016) totaled 101.

27

Figure 8. A heat map shows relatively even bird distribution with higher concentrations in the
middle and top portion of the runway.

28

Strikes by Year
165
149
138
115

113

113

117

92
76
52

52

30

2005

2006

2007

2008

2009

2010

2011

2012

2013

2014

2015

2016

Figure 2. SEA's strike database shows a rapid decline in reported strikes since installing the
FOD radar system.

29

Discussion and Further Research
The sheer volume of events captured by FODetect supports the claim of this
paper, that the hybrid system will reduce strikes through constant monitoring and quicker
response to live bird and wildlife FOD. While there are many variables to consider—
overall reporting, lower bird populations due to yearly variations, and improved
management—the system likely contributed to the lower strikes. This system also
provides greater insights about the level of bird and wildlife activity at SEA, which will
serve to guide future management decisions. While it is yet to be used for the latter, the
former was confirmed throughout the data analysis, which showed high bird (668) and
wildlife (4,072) counts.
Technical and Data Management Recommendations
The Xsight Systems FODetect Radar System drastically increases SEA’s runway
awareness and the wildlife team has a great amount of information readily available to
them. That being said, there are some areas where the system can be improved on the
user-side. Implementing standardized, detail-oriented operator reporting protocols would
greatly increase the clarity of the data. Throughout the study, the documentation feature
located in the Excel spreadsheet often included vague responses as well as mislabeled
fields (vegetation classified as wildlife) and included 300 null values. While using the
software the user can identify the objects using photos, but when working with the data
outside of FODetect, the level of detail is lacking. Improvements could be made with the
wildlife team working together with the airport operations crew to determine the best
criteria for describing what the system identifies in order to give a comprehensive and

30

useful description to work from. Creating a list of common terms and descriptors to
quickly enter into the description would expedite the data entry process.
While the FodDetect Radar System does include a camera, at this point, it is only
possible to export one image at a time. While working within the system or when looking
for known events the user has access to all images, but when using Excel or working
away from SEA, needing to access photos restricts work. However, it is possible to
develop code that would allow for mass photo exporting, which would improve data
analysis outside of FODetect’s software. If SEA plans to continue with researching and
using outside sources to help in the process, developing the ability to mass export photos
is recommended. A comprehensive study of the images would allow for a more detailed
the species on the runway and their abundance throughout the year.
Further Research
The temporal and geographic patterns related to bird and wildlife events identified
in this study can be used to guide further data exploration. Since this program is still in its
infancy, constant data collection and analysis in the future will allow for more specific
and fine-tuned research. Over time, a comprehensive, multi-year look at changes in
patterns in the radar data will be beneficial to the wildlife team. This may point to
consistent hotspots on the runway and allow for them to identify common species.
Combining the FOD radar with SEA’s other avian radar may also point to connections in
the flight and rest time around the airport. A more robust data set spanning several years
will also help spell out any long-term seasonal and daily patterns that may occur.
Looking at these patterns should allow the team to plan and execute more specific and
targeted management strategies.
31

Bird and wildlife events along the runway also have the potential to unlock
questions of relationships between the physical characteristics on and around the runway
and in the areas of high-density events. Monitoring and comparing wildlife events with
bird events may point to possible physical attractants along the runway and the
subsequent increase of insects or insect feeding birds. Merging the bird and wildlife
events by time may also help explain any wildlife bird attractants.
In terms of methodological improvements, developing a means for mass image
exporting would benefit further research to allow for species identification and
confirmation of bird and wildlife to aircraft strikes. Once this feature is available, an
analysis of strikes between aircraft and wildlife or birds and would allow the wildlife
team to determine the effectiveness of the system to harass birds and possibly limit
strikes in the future.

32

References
Anderson, C., & Osmek, S. (2005). Raptor strike avoidance at Seattle-Tacoma
International Airport: A biological approach. Bird Strike Committee-USA/Canada 7th
Annual Meeting.
Baker, J. A., & Brooks, R. J. (1981). Raptor and vole populations at an airport. Journal of
Wildlife Management, 45(1), 390-396.
Bayne, E. M., Scobie, C. A., & Rawson-Clark, M. (2012). Factors influencing the annual
risk of bird-window collisions at residential structures in Alberta, Canada. Wildlife
Research, 39(7), 583-592.
Belant, J. L., & Martin, J. A. (2011). Bird harassment, repellent, and deterrent
techniques for use on and near airports (Vol. 23). Transportation Research Board.
Belant, J. L., Washburn, B. E., & DeVault, T. L. (2013). Wildlife in airport
environments: Understanding animal movements at and near airports. USDA National
Wildlife Research Center - Staff Publications. Paper 1648.
Blackwell, B. F., DeVault, T. L., Fernandez-Juricic, E., & Dolbeer, R. A. (2009) Wildlife
collisions with aircraft: A missing component of land-use planning for airports.
Landscape and Urban Planning, 93(1), 1-9.
Blakwell, B. F., Schmidt, P. M., & Martin, J. A. (2013). Avian survey methods for use on
airports. In T. L. DeVault, B. F. Blackwell, & J. L. Belant (Eds.). Wildlife in airport
environments: Preventing animal-aircraft collisions through science-based management
(153-165). Baltimore, MD: Johns Hopkins University Press.
Buckley, P. A., & McCarthy, M. G. (1994). Insects, vegetation, and the control of
laughing gulls (larus articilla) at Kennedy International Airport, New York City. Journal
of Applied Ecology, 31(1), 291-302.
Cleary, E. C., Dolbeer, R. A., & Wright, S. E. (2004) Wildlife Strikes to civil aircraft in
the United States, 1990-2002. Federal Aviation Administration National Wildlife Strike
Database Serial Report Number 9.
Cleary, E. C., & Dolbeer, R. A. (2005). Wildlife hazard management at airports: A
manual for airport personnel. USDA National Wildlife Research Center – Staff
Publication. Paper 133.
Cleary, E. C., Dolbeer, R. A., & Wright, S. E. (2006). Wildlife strikes to civil aircraft in
the United States 1990-2005. Federal Aviation Administration National Wildlife Strike
Database Serial Report Number 12.

33

Coccon, F., Zucchetta, M., Bossi, G., Borrotti, M., Torricelli, P., & Franzoi, P. (2015). A
land use perspective for birdstrike risk assessment: The attraction risk index. PloS One,
10(6), e0128363. DOI: 10.1371/journal.pone.012836
Conomy, J. T., Collazo, J. A., Dubovsky, J. A., & Fleming, W. J. (1998). Dabbling duck
behavior and aircraft activity in coastal North Carolina. Journal of Wildlife Management,
62(1), 1127-1134.
Conover, M. R. (2002). Resolving human-wildlife conflicts. Boca Raton, FL: CRC Press.
DeVault, T. L., Belant, J. L., Blackwell, B. F., & Seamans, T. W. (2011). Interspecific
variation in wildlife hazards to aircraft: Implications for airport wildlife management.
2011 Bird Strike North America Conference.
Dolbeer, R. A. (2000). Birds and aircraft: Fighting for airspace in crowded skies. In
Proceedings of 19th Veterable Pest Conference (37-43). Davis, CA: University of
California Davis.
Dolbeer, R. A. (2006). Height distribution of birds recorded by collisions with civil
aircraft. Journal of Wildlife Management, 70(5), 1345-1350.
Dobleer, R. A. (2015). Trends in reporting of wildlife strikes with civil aircraft and in
identification of species struck under a primarily voluntary reporting system, 1990-2013.
Special report submitted to the Federal Aviation Administration. Retrieved from
https://www.faa.gov/airports/airport_safety/wildlife/media/trends-in-wildlife-strikereporting-1990-2013.pdf
Dolbeer, R. A., & Eschenfelder, P. (2003). Amplified bird strike risks related to
population increases of large birds in North America. Proceedings of the International
Bird Strike Committee, 26(1), 49-67.
Dolbeer, R. A., & Wright, S. A. (2009). Safety management systems: How useful will the
FAA National Wildlife Strike Database be? Human-Wildlife Conflicts, 3(2), 167-178.
Dolbeer, R. A., Wright, S. E., & Cleary, E. C. (2000) Ranking of hazard level of wildlife
species to aviation. Wildlife society Bulletin, 28(1), 372-378.
Dolbeer, R. A., Wright, S. E., Weller, J., & Begier, M. J. (2013). Wildlife strikes to civil
aircraft in the United States 1990-2012. Federal Aviation Administration National
Wildlife Strike Database Serial Report Number 19.
Dolbeer, R. A., Wright, S. E., Weller, J., & Begier, M. J. (2014). Wildlife strikes to civil
aircraft in the United States 1990-2013. Federal Aviation Administration National
Wildlife Strike Database Serial Report Number 20.

34

Federal Aviation Administration. (2009). Airport foreign object debris (FOD) detection
equipment. Advisory Circular. Retrieved from
https://www.faa.gov/documentLibrary/media/Advisory_Circular/150_5220_24.pdf
Federal Avation Administration. (2016). Aerospace forecast fiscal years: 2016-2036.
Retrieved from https://www.faa.gov/data_research/aviation/
Fretwell, S. D., & Lucas, H. L. (1970). On territorial behavior and other factors
influencing habitat distribution in birds. I. Theoretical development. Acta Biotheor, 19(1),
16-36.
Gauthreaux Jr, S. A., & Belser, C. G. (2003). Radar ornithology and biological
conservation. The Auk, 120(2), 266-277.
Gauthreaux, Jr, S. A., & Schmidt, P. M. (2013). Application of radar technology to
monitor hazardous birds at airports. In T. L. DeVault, B. F. Blackwell, & J. L. Belant
(Eds.). Wildlife in Airport Environments: Preventing Animal-Aircraft Collisions Through
Science-Based Management. Baltimore, MD: Johns Hopkins University Press.
Herricks, E. E., Woodworth, E., & Patterson, J. (2012). Performance assessment of a
hybrid radar and electro-optical foreign object debris detection system. Report to Federal
Aviation Administration. Retrieved from http://www.tc.faa.gov/its/worldpac/techrpt/tc1222.pdf
Klem, D. (1990). Collisions between birds and windows: Morality and prevention.
Journal of Field Ornithology, 6(1), 120-128.
Klem, D. (2010). Sheet glass as a principal human-associated avian morality factor. In S.
K. Majumdar, T. L. Master, M. Brittingham, R. M. Ross, R. Mulvihill, & J. Huffman
(Eds.), Avian ecology and conservation: A Pennsylvania focus with national implications
(Chapter 20). Easton, PA: Pennsylvania Academy of Science.
Krebs, C. J. (2001). Ecology - The experimental analysis of distribution and abundance.
San Francisco, CA: Benjamin Cummings.
Linnell, M. A., Conover, M. R., & Ohashi, T. J. (1996). Analysis of bird strikes at a
tropical airport. Journal of Wildlife Management, 60(1), 935-945.
Linnell, M. A., Conover, M. R., & Ohashi, T. J. (1999). Biases in bird strike statistics
based on pilot reports. Journal of Wildlife Management, 63(1), 997-1003.
National Transportation Safety Board. (2010, May 04). Loss of thrust in both engines, US
Airways flight 1549 Airbus Industrie A320-214, N106US. Investigations Accident Report
Detail. Retrieved from
http://www.ntsb.gov/investigations/AccidentReports/Pages/AAR1003.aspx

35

Osmek, S. (Month, 2016). Personal communication.
Port of Seattle. (2015). Airport Statistics. Retrieved from
https://www.portseattle.org/About/Publications/Statistics/AirportStatistics/Pages/default.aspx
Port of Seattle. (n.d.). Sea-Tac Airport's comprehensive wildlife management programs.
Retrieved from http://www.portseattle.org/Environmental/Water-WetlandsWildlife/Pages/Wildlife-Management.aspx
Port of Seattle, Seattle-Tacoma International Airport, U.S. Department of Agriculture
Animal and Plant Health Inspection Service, & Wildlife Services. (2004). Wildlife hazard
management plan: Maximize safety now… Minimize future risk. SEA airport
certification manual. Retrieved from https://www.portseattle.org/Environmental/WaterWetlands-Wildlife/Documents/wildlifemanagement08.pdf
Reiter, D., Brunson, M., & Schmidt, R. H. (1999). Public attitudes toward wildlife
damage management and policy. Wildlife Society Bulletin, 28(4), 931-758.
Seattle Port Commission. (2012). Century agenda. Retrieved from
https://www.portseattle.org/About/Commission/Pages/Century-Agenda.aspx
Solman, V. E. F. (1981). Birds and aviation. Environmental conservation 8(1), 45-51.
Thorpe, J. (2003). Fatalities and destroyed aircraft due to bird strikes, 1912-2002.
Proceedings of the International Bird Strike Committee, 26(1), 85-113.
Yonemoto, N., Kohmura, A., Futatsumori, S., Uebo, T., & Saillard, A. (2009, October).
Broad band RF module of millimeter wave radar network for airport FOD detection
system. In Radar Conference-Surveillance for a Safer World, 2009. RADAR.
International (pp. 1-4). IEEE.

36

Appendices

Appendix 3. Map showing the 101 hybrid surface detection units along the center runway
at SEA. Each unit is placed roughly 200 feet apart vertically and 150 feet across the
runway.

37

Appendix 4. The FODetect process involves both the autonomous hybrid camera/radar
system as well as human operators. If the unit detects FOD, an alert is created and the
operator can take control of the unit to determine the threat level and any necessary
actions. Image: Xsights

38

Appendix 5. A screen grab from the Xsights software showing a strike. The red rectangle
around the struck bird highlights the object as detected by the automated system.

Appendix 4. One of the many mantises spotted by the FODetect system. The rise in
mantises coincided with a decrease and overall low value of grasshoppers.

39

Appendix 5. A worm discovered by the FODetect system. Worms were often a precursor
to increased bird activity, specifically gulls.

40