University of Southern California Dissertations and Theses

Improving wetland determination utilizing unmanned aerial systems

(USC Thesis Other)

Improving wetland determination utilizing unmanned aerial systems

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

Improving Wetland Determination Utilizing Unmanned Aerial Systems

by

Monika Burchette

A Thesis Presented to the
Faculty of the USC Graduate School
University of Southern California
In Partial Fulfillment of the
Requirements for the Degree
Master of Science
(Geographic Information Science and Technology)

December 2018

ii

Copyright ® 2018 by Monika Burchette

iii

This paper is dedicated to my mother and father, Lori and Michael Burchette. Without their
unending encouragement and faith in my abilities I would have never attempted this program.

iv

Acknowledgements
I am grateful to my mentor, Dr. John Wilson, for his enthusiasm and the direction he has
provided for this thesis, and my other thesis guidance committee members, Drs. Andrew Marx
and Travis Longcore, whose insight and expertise were invaluable during the review process. I
would also like to thank my employer, Olsson Associates, for allowing me to explore unmanned
aerial systems and allowing me the use of company software to complete this thesis. Specifically,
I would like to extend thanks to my supervisors, Reza Khakpour, and William “Buck” Ray, as
well as Project Managers, Amanda Miller and Hilary Clark. They each allowed me to talk
through ideas, develop explorable notions and supported my journey throughout the project.
A special thank you to all the landowners that allowed me the access to their properties:
the Hillis Family, Michael and Lori Burchette, and the Davis Family. You all made this project
possible and without your permission this would have remained just an idea. Thank you to all of
my supportive Oklahomans! Finally, I would also like to thank my fiancé, Zac Wheat, for putting
up with me during some of the more stressful times of this process and being a steady supportive
presence in my life.

v

Table of Contents
Acknowledgements ........................................................................................................................ iv
List of Figures ............................................................................................................................... vii
List of Tables ................................................................................................................................. ix
List of Abbreviations ...................................................................................................................... x
Abstract .......................................................................................................................................... xi
Chapter 1: Introduction ................................................................................................................... 1
1.1 History of the Clean Water Act ...........................................................................................2
1.2 Traditional Methods of Wetland Determination and Delineation .......................................5
1.3 Unmanned Aerial Systems ...................................................................................................7
1.4 Motivation and Goals ...........................................................................................................8
1.5 Thesis Organization .............................................................................................................9
Chapter 2: Related Work .............................................................................................................. 10
2.1 Past Wetland Studies Using Remotely Sensed Imagery ....................................................10
2.2 Current Methods of Wetland Determination .....................................................................13
2.3 The Use of UAS for the Collection of Spatial Data ...........................................................15
Chapter 3: Methodology ............................................................................................................... 19
3.1 Case Studies ..................................................................................................................19
3.2 Equipment .....................................................................................................................20
3.2.1 Systems and Software ...............................................................................................20
3.2.2 Trimble GPS Receiver ..............................................................................................25
3.2.3 Software systems ..................................................................................................25
3.3 Data Acquisition ............................................................................................................26
3.3.1 UAS Data Acquisition ..............................................................................................26
3.3.2 GPS Data Acquisition and Wetland delineation .......................................................27
3.3.3 Esri Base Maps and Google Earth Imagery Acquisition......................................28
3.4 Post-Processing Data .....................................................................................................28
3.4.1 UAS Post-Processing ................................................................................................28
3.4.2 GPS Post-Processing .................................................................................................29
3.5 Data Analysis ................................................................................................................29
vi

Chapter 4: Results ......................................................................................................................... 32
4.1 Accuracy of UAS and Existing data ..................................................................................32
4.2 Comparison of NWI, Delineation, and UAS Collection Methods ................................38
4.2.1 Case Study #1 ...........................................................................................................39
4.2.2 Case Study #2 ...........................................................................................................41
4.2.3. Case Study #3 ..........................................................................................................43
4.2.4 Case Study #4 ...........................................................................................................45
4.2.5 Case Study #5 ...........................................................................................................47
4.2.6. Case Study #6 ..........................................................................................................49
4.2.7 Case Study #7 ...........................................................................................................51
4.2.8 Case Study #8 ...........................................................................................................53
4.2.9 Case Study #9 ...........................................................................................................55
4.2.10 Case Study #10 .......................................................................................................57
4.2.11 Case Study #11 .......................................................................................................59
4.2.12. Case Study #12 ......................................................................................................61
4.2.13. Case Study #13 ......................................................................................................63
4.2.14. Case Study #14 ......................................................................................................65
4.2.15. Case Study #15 ......................................................................................................67
4.3 Cost-Benefit Analysis ........................................................................................................69
Chapter 5: Discussion and Conclusions ........................................................................................ 70
5.1 Findings ..............................................................................................................................70
5.2 Advantages and Disadvantages of Using the UAS Methodology .....................................71
5.2.1 Supplemental buffers ................................................................................................72
5.2.2 Seasonality ................................................................................................................73
5.2.3 Other potential methods ............................................................................................79
5.3 Conclusions ........................................................................................................................81
REFERENCES ............................................................................................................................. 82

vii

List of Figures
Figure 1. Google Earth images versus Esri base maps ................................................................. 15
Figure 2. Esri base map with NWI mapped wetland overlaid on top of the basemap .................. 15
Figure 3. Phantom 3 Pro UAS by DJI........................................................................................... 22
Figure 4. Pro gimbal camera roll arm and camera ........................................................................ 23
Figure 5. Example of pitch, roll and yaw on UAS ........................................................................ 23
Figure 6. Example of Pix4Dcapture mission ................................................................................ 24
Figure 7. Oklahoma NAD 1983 UTM zones ................................................................................ 30
Figure 8. Oklahoma Annual Precipitation from 1895-2017 ......................................................... 33
Figure 9. Oklahoma Annual Temperature from 1895-2017 ......................................................... 34
Figure 10. Case Study # 1 in Meeker, OK .................................................................................... 40
Figure 11. Case Study # 2 in Meeker, OK .................................................................................... 42
Figure 12. Case Study # 3 in Meeker, OK .................................................................................... 44
Figure 13. Case Study # 4 in Meeker, OK .................................................................................... 46
Figure 14. Case Study # 5 in Pawnee, OK .................................................................................... 48
Figure 15. Case Study # 6 in Roosevelt, OK ................................................................................ 50
Figure 16. Case Study # 7 in Roosevelt, OK ................................................................................ 52
Figure 17. Case Study # 8 in Roosevelt, OK ................................................................................ 54
Figure 18. Case Study # 9 in Roosevelt, OK ................................................................................ 56
Figure 19. Case Study # 10 in Roosevelt, OK .............................................................................. 58
Figure 20. Case Study # 11 in Roosevelt, OK .............................................................................. 60
Figure 21. Case Study # 12 in Roosevelt, OK .............................................................................. 62
Figure 22. Case Study # 13 in Roosevelt, OK .............................................................................. 64
Figure 23. Case Study # 14 in Roosevelt, OK .............................................................................. 66
Figure 24. Case Study # 15 in Roosevelt, OK…………………………………………………...66
viii

Figure 25. A map of Case Study #11 showing how the application of a buffer around the UAS
mapped wetlands allowed for the delineated wetland to be fully encompassed within a 15 ft
buffer ..................................................................................................................................... 76
Figure 26. A map of Case Study #7 show how the application of a buffer around the UAS
mapped wetlands which does not consider the topography with a 35 ft buffer .................... 77
Figure 27. Differences in overlap between winter and summer UAS flight times and the
delineated wetland for Case Study #5 ................................................................................... 78
Figure 28. Slope estimate from Digital Surface Model provided in Pix4D with current wetland
extent ..................................................................................................................................... 80
Figure 29. Wetland growth estimate from Digital Surface Model provided in Pix4D with current
wetland extent ........................................................................................................................ 80

ix

List of Tables
Table 1. List of field equipment and specifications ...................................................................... 21
Table 2. Capture dates for imagery data in Meeker, OK case studies .......................................... 32
Table 3. Capture dates for imagery data in Pawnee, OK case study ............................................ 32
Table 4. Capture dates for imagery data in Roosevelt, OK case studies ...................................... 32
Table 5. The average maximum, mean, and minimum temperatures and rainfall totals for
Meeker, OK three months prior to imagery capture dates .................................................... 34
Table 6. The average maximum, mean, and minimum temperatures and rainfall totals for
Pawnee, OK three months prior to imagery capture dates .................................................... 35
Table 7. The average maximum, mean, and minimum temperatures and rainfall totals for
Roosevelt, OK three months prior to imagery capture dates ................................................ 35
Table 8. NWI mapped wetlands attributes .................................................................................... 36
Table 9. Acreages calculated from NWI, Delineation, and UAS collection methods .................. 38
Table 10. Time required to acquire, process and analyze data collected for delineation and UAS.
............................................................................................................................................... 69
Table 11. Percent overlap achieved using at 5, 10, 15, 20, 25, 30, and 35 ft buffers generated
around the 15 UAS mapped wetlands ................................................................................... 74

x

List of Abbreviations
AVHRR Advanced Very High-Resolution Radiometer
CASI Compact Airborne Spectrographic Imagery
CIR Color-infrared
FEMA Federal Emergency Management Agency
GeoEye Geologic Event Volume Estimator
GPS Global Positioning System
MODIS Moderate Resolution Imaging Spectroradiometer
NHD National Hydrology Dataset
NPDES National Pollutant Discharge Elimination Systems
NRCS National Resources Conservation Services
NWI National Wetland Inventory
RGB Red Green Blue
SWOT Strength Weakness Opportunity, and Threats
UAS Unmanned Aerial Systems
USACE United States Army Corp of Engineer
USCWA United State Clean Water Act
USEPA United States Environmental Protection Agency
USFWS United States Fish and Wildlife Service
WOTUS Waters of the United States

xi

Abstract
When project proponents wish to assess a development site for jurisdictional wetland impacts,
they are traditionally left with two options: a wetland determination or a delineation. A wetland
determination is customarily a desktop assessment of the site including, but not limited to, the
following datasets: National Wetland Inventory (NWI), National Hydrography Dataset (NHD),
National Resources Conservation Services Soil Survey (NRCS), topographic maps and satellite
imagery. A wetland delineation assesses the presence of hydrophytic vegetation, hydric soils and
hydrology during field evaluation. The NWI is typically used to determine where existing
wetlands are in order to determine if they qualify as jurisdictional wetlands. This allows project
proponents to either take the appropriate avoidance measures to reduce impacts to the wetland or
determine if a full wetland delineation is required to apply for a Section 404 permit. In some
cases, NWI maps have not been updated for up to 30 years, and these mapped wetlands are
limited by conditions that were present at the time the aerial imagery was taken.
This thesis shows that by incorporating unmanned aerial systems (UAS) into a wetlands
determination, wetland specialists and project planners can capture current conditions of the
development site (i.e. topography, disturbance, land cover, etc.) within efficient time frames and
assess the potential extent of a wetland(s). This allows project proponents to avoid the cost and
time restrictions that come from a full wetland delineation. The UAS imagery was compared to
historically mapped wetlands still present; UAS improved placement of wetlands on the
landscape and had on average a 76.5% overlap with delineated wetlands. Future research into
buffer distances, topography, seasonality and thermal imagery could improve this overlap. With
the aerial imagery from current conditions, wetland specialists can assess potential wetland
extent, hydrology, highwater marks, and coarsely classify vegetative condition.
1

Chapter 1 : Introduction
An increasingly important issue when developing lands is wetland assessment and
avoidance. Wetlands are valued as important sinks, sources, and transformers of a multitude of
chemical and biological materials, as well as being valued as vital habitat for many fish and
wildlife species. Wetlands have been described by scientists as the “kidneys of the landscape”.
This phrase is used because wetlands typically act as downstream filters of water and both
natural and human-made waste. Wetland areas are known to provide ecosystem services such as
water purification, sediment and nutrient retention, groundwater replenishment and flood control
(Zhang et al. 2010) which is vital to many ecosystems and human activities.
Prior to the mid-1970s, wetlands were considered to be areas where insect-born diseases
such as malaria originated. Therefore, the drainage and destruction of wetlands was encouraged,
with states like California and Ohio reporting up to a 90% loss in natural wetlands. Once the
valuable services of wetland areas were recognized, the U.S. government supported a multitude
of federal, state and private programs to preserve existing wetlands, placing a higher priority on
wetlands connected to hydrology. In the U.S., the U.S. Fish and Wildlife Service (USFWS) is
involved with the classification and inventory of wetlands, the U.S. Environment Protection
Agency (USEPA) is involved with human activity and its impacts around wetlands, and the U.S.
Army Corp of Engineers (USACE) provides guidelines to assess wetlands on the ground and
gives permits for wetland destruction. The USEPA reserves a veto authority over permits given.
As defined by the USACE and the USEPA, "Wetlands are areas that are inundated or
saturated by surface or ground water at a frequency and duration sufficient to support, and that
under normal circumstances do support, a prevalence of vegetation typically adapted for life in
saturated soil conditions. Wetlands generally include swamps, marshes, bogs, and similar areas"
2

(USEPA 1970). While some states have more stringent laws regarding wetlands, this thesis
focuses on the federal guidelines established by the U.S. Clean Water Act (USCWA).
1.1 History of the Clean Water Act
In 1899 the River and Harbors Act was passed by the U.S. Congress to regulate the
placement of anything that might affect navigation in navigable waters; this includes the
construction of any bridge, dam, dike, wharf, pier, dolphin, boom, weir, breakwater, bulkhead,
jetty or causeway (33 U.S.C. 403; Chapter 425, March 3, 1899; 30 Stat. 1151). It also stated that
it shall not be lawful to excavate or fill, or in any manner to alter or modify the course, location,
condition, or capacity of, any port, roadstead, haven, harbor, canal, lake, harbor of refuge, or
enclosure within the limits of any breakwater, or of the channel of any navigable water of the
U.S., unless the work has been recommended by the Chief of Engineers and authorized by the
Secretary of War prior to beginning the same. In 1977 the USCWA was implemented to revise
the River and Harbors Act. This included Section 401, which deals with water quality
certification; Section 402 which deals with National Pollutant Discharge Elimination Systems
(NPDES) or liquid discharge; and Section 404, which deals with placement of fills in Waters of
the U.S. (WOTUS). The term WOTUS means: all waters which are currently used, or were used
in the past, or may be susceptible to use in interstate or foreign commerce; all interstate waters
including interstate wetlands; all other waters such as intrastate lakes, rivers, streams (including
intermittent streams), mudflats, sandflats, wetlands, sloughs, prairie potholes, wet meadows,
playa lakes, or natural ponds, the use, degradation or destruction of which could affect interstate
or foreign commerce (USEPA 2017).
Wetlands that flow into navigable WOTUS are known as jurisdictional wetlands, which
are the wetlands that are typically permitted during development. Section 404 of the USCWA
3

establishes a program to regulate the discharge of dredged or fill material into WOTUS,
including wetlands. A Section 404 permit, whether individual or general, allows for a project
proponent to discharge dredge or fill materials into the WOTUS during development of a
specified area within the limitations set forth for each permit. In order to comply with Section
404 of the USCWA and associated regulations, project proponents must assess the current
conditions of each proposed development site before applying for an individual Section 404
permit. This can prolong project timelines and generate uncertainty for stakeholders.
Without going to the project site, many project proponents rely on wetland
determinations to remotely gather data regarding the project site, which can vary in accuracy and
may have long gaps between collection and development dates. For example, some NWI
wetlands mapped by the USFWS in Oklahoma have not been updated since September 1981.
The methods associated with determinations are known to have flaws in location and size, with
the USFWS stating the following on their website for the NWI: “The Service's objective of
mapping wetlands and deep-water habitats is to produce reconnaissance level information on the
location, type and size of these resources. The maps are prepared from the analysis of high
altitude imagery. Wetlands are identified based on vegetation, visible hydrology and geography.
A margin of error is inherent in the use of imagery; thus, detailed on-the-ground inspection of
any particular site may result in revision of the wetland boundaries or classification established
through image analysis. The accuracy of image interpretation depends on the quality of the
imagery, the experience of the image analysts, the amount and quality of the collateral data, and
the amount of ground truth verification work conducted. Metadata should be consulted to
determine the date of the source imagery used and any mapping problems. Wetlands or other
mapped features may have changed since the date of the imagery and/or field work. There may
4

be occasional differences in polygon boundaries or classifications between the information
depicted on the map and the actual conditions on site” (USFWS 2018a).
In order to improve accuracy, wetland scientists in the past have used publicly available
aerial imagery from Google and Esri to determine wetland location but are limited by the
collection date and the resolution of the imagery provided (Scarpace et al. 1982). On the other
hand, a wetland scientist can go out onsite and visually assess the wetland using parameters put
in place by the USACE (1987) wetland delineation manual along with applicable regional
supplements. A wetland delineation is a costly and time-consuming effort that requires hours of
survey time and assessment to meet the manual’s guidelines. It also requires a specialist that has
been certified as a wetland delineator in order for the assessment to be considered as valid by the
USACE.
In recent years, many studies have used UASs for surveying and monitoring natural
landscapes, such as forests and rangeland (e.g. Hardin and Jackson 2005; Dunford et al. 2009;
Rango et al. 2009; Breckenridge and Dakins 2011; Getzin et al. 2012). Benefits stated from these
past studies include the ability to acquire aerial imagery at very high spatial resolution (<10
cm/pixel), and the ability to deploy the UAS in a convenient, timely and repeatable manner that
is cost competitive when compared to traditional survey methods (Chabot and Bird 2013). The
utilization of UAS can also decrease harm and injury to researchers by allowing a researcher to
assess the landscape without encountering the natural hazards which may be encountered when
in the field.

5

1.2 Traditional Methods of Wetland Determination and Delineation
Effective conservation and management of wetlands depends on the ability to collect
accurate and timely data on the habitats that contain them (Finlayson and Mitchell 1999).
Traditionally, project proponents have been left with a choice between a wetlands determination
or a complete wetlands delineation in order to assess wetlands during project development and
determine the appropriate permitting pathway under Section 404 of USCWA.
For the purposes of this thesis, a wetlands determination will be defined as the study of
remotely sensed data acquired from the NWI, NHD, NRCS Web Soil Survey, Federal
Emergency Management Agency (FEMA) flood maps and aerial imagery (sourced by one or
more of the following: Esri, DigitalGlobe, GeoEye, i-cubed, U.S. Department of Agriculture –
Farm Service Agency, U.S. Geological Survey, AEX, Getmapping, Aerogrid, Institut
Geographique National, Interior Gateway Protocols, Swisstopo, or the GIS User Community).
These datasets are used to determine where historic wetlands are located within a development
area, as well as to determine if those historical wetlands are potentially jurisdictional wetlands. A
wetlands determination allows project proponents to determine what wetlands have been
historically present within a development area; however, a wetlands determination is not legally
binding for project proponents and does not meet the standards set forth by the USACE for
acquisition of a Section 404 permit.
Current wetlands determinations are limited by the accuracy and precision of the above-
mentioned historic datasets. For example, NWI features may not capture the size and location of
the associated wetlands and some areas of the country have not been mapped by FEMA, so
flooding conditions are unknown. Many NWI maps in Oklahoma are 30 years old (OKGOV
2018) and the satellites used coarse resolutions of 10 m or greater. Further, wetland
6

determinations are also limited by the environmental conditions and time frame in which the
aerial imagery was captured. Consultants are unable to control the season, time-of-day, or
resolution in which the images are captured (Lee and Lunetta 1995; Adam et al. 2010). This
inaccuracy contributes to unknown errors of wetland placement and until a site visit occurs the
wetland feature’s status is not known.
In contrast, a formal wetlands delineation involves an onsite assessment of three wetland
indicators as defined by the USACE’ Wetlands Delineation Manual (USACE 2010):
1. Hydric soil - soils saturated, flooded, or ponded, long enough during the growing
season to develop anaerobic conditions in the upper profile.
2. Hydrology – presence of water (past or present).
3. Hydrophilic vegetation - plant life growing in water, soil, or on a substrate that is
periodically deficient in oxygen due to excess water.
Presence or absence of all three indicators is captured by the wetlands delineator to
determine the extent of the wetland. To determine the wetland boundary, the delineator chooses a
series of data points that are representative of the site. The delineator digs a soil pit at each data
point in sample wetland areas and sample upland areas for vegetation and hydrology. In some
regions, wetland delineations are limited to being performed during the region’s growing season
for delineators to fully assess vegetation characteristics. Wetland delineations are used by the
USACE to assess the jurisdictional status of delineated wetlands, identify the boundaries of the
delineated wetlands and are legally binding for the project proponent. However, wetlands are
challenging to survey because of their often complex patchwork of flooded areas interspersed
with dense vegetation, which can be laborious for a wetlands delineator to navigate and
characterize at ground level. Due to the large scale of some development projects, the costs and
7

time associated with performing a formal wetlands delineation might impede on the planning and
progression of the project. Currently wetland delineation is considered to be the most accurate
form of wetland assessment.
1.3 Unmanned Aerial Systems
Recent developments in UAS platforms, positional and attitudinal measurement sensors,
imaging sensors, and processing approaches have opened a vast new area of opportunities in
remote sensing for observation, measurement, mapping, monitoring, and management in various
natural environments (e.g. forests). UAS are an ideal tool for monitoring sensitive areas and
subjects that may be threatened or destroyed if monitored manually (Jones IV, Pearlstine, and
Percival 2006). The use of high spatial resolution aerial imagery captured by small UAS in
natural resource management is rapidly increasing (Abd-Elrahman, Pearlstine, & Percival, 2005;
Laliberte, Rango, & Herrick, 2007; Rango et al., 2006; Watts et al., 2012). Improvements in
technology and procedures are gradually enabling UASs to produce high-quality georeferenced
orthorectified images through software programs like Pix4D. The spatial and temporal
resolutions of UAS imagery are controlled by the operator/user who determines the mission
parameters (e.g. time of day, season, resolution, percent overlap and flying height); this gives a
significant advantage, even over traditional piloted image-capture missions. Being able to assess
a landscape under controlled conditions allows for repeatable flights to be conducted over the
same area in order to see changes over time. This has been accomplished in the past with other
historic satellite imagery in order to assess vegetation growth or contraction (Everitt et al. 2010).
Although more sophisticated image capturing sensors have been created (i.e. hyperspectral and
LiDAR sensors) for smaller UASs these sensors are very expensive. Off-the-shelf three-band
(RGB) cameras can provide the resolution needed for many of the current studies. This also sets
8

a baseline for data capture that is affordable for many researchers. Inexpensive alternatives are
best when considering the likelihood of a UAS failure or crash. Many over-the-counter units can
be purchased for less than $1,000, which makes start-up cost for any company wishing to use
UAS reasonable to maintain, and replacement costs are negligible in cases of unit failure. While
the UAS model would not be meant to replace a full wetland delineation, it could be used as a
planning and avoidance tool for project development. The ability to acquire up-to-date data for a
project area as well as a visual from the air which could be used to assess the approximate
location and size of the wetland. With some USACE districts, aerial photos are required with the
submission for a permit; however, due to the time gaps in image capture and project dates the
USACE may not accept aerial photos or satellite images in lieu of a full delineation. If UAS
imagery could be utilized as an alternative to historic images both project proponents and
USACE would have the opportunity to use the most current data for the assessment of the
Project Area.
1.4 Motivation and Goals
Wetlands are highly protected sites that require detailed information in order to
accurately assess the conditions currently present. Historical datasets are by nature inaccurate
and make current wetland determinations limited in the information that is being provided. This
leaves project proponents with little to no preplanning of site development until a full wetland
delineation has occurred to assess the size and shape of potentially present wetlands. UASs have
been utilized in many environmental studies to capture current conditions in a timely and cost-
effective manner. The goals of this study were to: (1) improve upon current wetland
determination datasets; (2) determine the accuracy of the imagery when compared to mapped
9

NWI wetlands and wetland delineations; and (3) determine the cost effectiveness of UAS when
compared to traditional methods.
1.5 Thesis Organization
The next chapter discusses previous studies from the 1980s to the present (Section 2.1),
current methods of wetland determination and delineation used as a standard for environmental
consulting agencies (Section 2.2) and methods of UAS data collection (Section 2.3). Chapter 3
details the collection, processing and analysis methods used for this thesis. Chapter 4 describes
the results that were developed for the case study wetlands, and Chapter 5 discusses conclusions
and future research questions.

10

Chapter 2 : Related Work
Remote sensing of aerial imagery collected by satellites has been utilized in a number of
environmental research projects involving wetlands. UAS have also become more prominent in
environmental assessments. This chapter will review past studies of remotely sensed wetlands
(Section 2.1), current methods used to conduct wetland determinations (Section 2.2), and the
adoption of UAS for the collection of spatial data (Section 2.3).
2.1 Past Wetland Studies Using Remotely Sensed Imagery
The development of satellite remote sensing in the 1970s made researchers begin to
consider using remote sensing for wetland analyses. However, only 17 papers were found to be
published before the 1990s when technological advancements began to boom. Johannessen
(1964) utilized aerial photography of Nehalem Bay from 1939 to 1960 to compare marsh
boundaries. Johannessen recognized a circular pattern on the photos that he determined to be
clumps of vegetation on mud flats that proved rapid expansion of the marshes. This is the earliest
known wetland remote sensing paper. In the following years, wetland mapping and visual
interpretation was developed for aerial imagery; however, this approach proved to be difficult for
past researchers to use due to the inability to accurately identify wetland cover classes and plants.
Scarpace et al. (1982) used digitized aerial photography to identify the wetland vegetation in
marshlands by employing visual interpretation methods, with an accuracy that was determined to
be 56-60%.
Color-infrared (CIR) aerial photographs were found to clearly identify vegetation types,
with strongly reflective plants in near-infrared wavebands (Dale et al. 1986). Lovvorn and
Kirkpatrick (1982) found that CIR photos at the 1:4800 scale can identify dominant plant species
and that early September in Indiana was an optimal time for identifying dominant species
11

because of the senescence of species in the fall. Tiner (1990) used high-altitude aerial
photography with scales ranging from 1:40,000 to 1:130,000 as the primary data sources and
visual stereoscopic photo interpretation for the identification, classification, and inventory of
forested wetlands on a national basis in the U.S. This author concluded that CIR aerial
photography from the early spring is best for detecting deciduous forested wetlands in
temperate regions.
Besides the visual interpretation techniques for wetland mapping, automated
unsupervised and supervised classification methods were also utilized. Supervised
classification was used for high-resolution multispectral imagery from the Compact Airborne
Spectrographic Imager (CASI) to classify mangroves (Green and Ellis 1998). The identified
mangroves had an overall accuracy of 78.2%. The authors concluded that CASI imagery can
be used to assess mangrove areas with a greater level of detail and accuracy than with satellite
sensors. Unsupervised and supervised image analysis techniques were used for archived aerial
CIR photographs to monitor black mangrove on the South Texas Gulf Coast of the U.S. and
provided accurate results (Everitt et al. 2014, 2015).
Aerial photography has an advantage in terms of spatial resolution and data acquisition
time; however, wetland studies use these data typically in narrow coastal areas or along rivers
because of the generally small areas covered by these photos. In most cases, aerial
photographs are combined with other satellite images to study wetlands on a regional or
national scale. Aerial surveys can also be used as supplementary data for the land cover
mapping in large areas based on coarse spatial resolution imagery (e.g., Advanced Very High
Resolution Radiometer, AVHRR) (Rogers et al. 1997). Aerial photography techniques have
been widely used in wetland studies because of their excellent advantages in terms of spatial
12

resolution, cost and time, especially when satellite remote sensing techniques were relatively
new. Because of the challenges of data acquisition for large areas by flight, aerial photography
has usually been used for the mapping of small wetland areas. After the launch of satellites,
especially Landsat TM, aerial photography was mainly used for assessment of the classification
procedures or biomass derived from lower-resolution remote sensing methods. Satellite remote
sensing data provides an effective and efficient tool for detecting water body areas and flood
inundation extent over large areas. Because of the high temporal resolution and large coverage,
Moderate Resolution Imaging Spectroradiometer (MODIS) has significant advantages for
mapping the wetland extent and dynamics at a coarse spatial resolution (Ordoyne and Friedl
2008). Cai et al. (2005) used MODIS data to map the water body areas of Poyang Lake in China
and obtained the lake surface area with an error of approximately ± 6.19%.
High spatial resolution is considered to be images with the spatial resolution of <4 m,
including data from SPOT-5, IKONOS, Quickbird, WorldView, and GeoEye. Compared with
medium-resolution and hyperspectral images, images with high spatial resolution have more
geometry and texture information on the surface features and can be used to identify ground
features more easily (Guo et al. 2017). Images with high spatial resolution provide detailed
information about the ground surface and are a cornerstone of remote sensing. Because of the
high price of high-resolution images (Lee and Lunnetta 1995), they are mainly used in small
study areas, to explore new methods, or to verify the wetland map accuracy. Many researchers
have confirmed that high resolution images have the potential to improve wetland
classification accuracy (e.g. Guo et al. 2017).

13

2.2 Current Methods of Wetland Determination
Currently a wetland determination is deemed a routine or minimum-level of wetland assessment.
This is because a consultant or wetland scientist can conduct a “reconnaissance” and characterize
the scope of the work. A determination indicates where potential jurisdictional or general
wetlands may be located. It involves a desktop evaluation of NWI, NHD, NRCS Soil Survey,
topographic maps and satellite imagery. The following is a description of the datasets:
1. NWI - publicly available resource that provides detailed information on the abundance,
characteristics, and distribution of U.S. wetlands. NWI data are used by natural resource
managers, within the USFWS and throughout the nation, to promote the understanding,
conservation and restoration of wetlands.
2. NHD - represents the nation’s drainage networks and related features, including rivers,
streams, canals, lakes, ponds, glaciers, coastlines, dams, and stream gages.
3. NRCS Web Soil Survey - provides soil data and information produced by the National
Cooperative Soil Survey. NRCS has soil maps and data available online for more than 95
percent of the nation’s counties and anticipates having 100 percent in the near future.
4. Topographic maps - general use maps at medium scales that present elevation (contour
lines), hydrography, geographic place names, and a variety of cultural features. Current-
generation topographic maps are created from digital GIS databases, and are branded "US
Topo."
5. Satellite imagery – Esri base maps are typically used for these reports and includes map
features on 0.3 m resolution imagery in the continental U.S. and 0.6 m resolution imagery
in parts of western Europe from Digital Globe. In other parts of the world, 1 m resolution
imagery is available from GeoEye IKONOS, Getmapping, AeroGRID, and IGP Portugal.
14

One meter USDA NAIP imagery is available in some states of the U.S. Additionally,
imagery at different resolutions has been contributed by the GIS user community.
6. LiDAR Data - While LiDAR is a more accurate tool, it is cost prohibitive as many
LiDAR devices cost up to $10,000 and the state of Oklahoma does not have publicly
available high-resolution data at this time.
After these datasets are considered, the scientist determines existing wetlands and
“potential jurisdiction wetlands”. Potential jurisdictional wetlands would be an area that exhibits
one or more of the three wetland criteria found in a wetland delineation. These features allow a
project proponent to appreciate the conditions and choose if a Section 404 permit is necessary or
can be avoided (Lyon and Green1992). If the client determines that wetlands will need to be
assessed before development then, the only option to them is to hire a wetland delineator to do an
onsite assessment of the site.
There are two main problems with the wetland determination. One is that the data is not
likely to represent current conditions. Figure 1 shows a sample wetland, using Google Earth Pro
with a 30 cm per pixel resolution but image taken on 02/25/2014 and Esri base maps with a 0.5
m per pixel resolution and imagery that was collected on 04/01/2016. If a researcher is looking to
assess a wetland they are left with a choice between the two images: is the resolution or the date
that is the most current and therefore more likely to represent any modifications that have
occurred more important? The second flaw comes from the NWI maps, because years or even
decades may have passed from the date of assessment compared to the original date the map was
prepared. The associated wetland that was mapped in Figure 2 was mapped in September, 1981,
which when this thesis was written would be a 37-year difference in dates. It is unlikely that a
wetland that was mapped 37 years ago would represent current conditions on the site.
15

Figure 1. Google Earth images versus Esri base maps

Figure 2. Esri base map with NWI mapped wetland overlaid on top of the basemap.
2.3 The Use of UAS for the Collection of Spatial Data
The process of capturing aerial imagery using a manned aircraft is a time consuming and
costly endeavor. Project proponents will incur the costs of hiring an aircraft, pilot, and insurance.
16

They will also need to plan the missions (i.e. study areas, buffers, transect distances), plan a
flight time, determining the optimal resolution of the imagery, and obtaining the necessary
photographic equipment (Falkner and Morgan 2002). The correct weather conditions for flights
are also considered, if unfavorable conditions such as fog, high winds, or rain appear on the day
of the flight, pilots and scientists will be grounded until favorable conditions occur. Even though
no flight has taken place many aviation companies will still charge a daily fee for having the
vehicle on standby. This leads to unneeded losses in time and money. No matter which vehicle
(fixed-wing, helicopter, or UAS) is being used to capture aerial images, mission planning
processes are similar for proper image acquisition, with the main differences being the height
and speed at which the missions are flown.
Paine and Kiser (2012) provides an excellent list of variables that need to be addressed
before any aerial photography mission. The mission will incorporate the altitude that must be
maintained, the percent overlap of the images, the pattern that will be flown, the angle of the
camera, the appropriate focal length, and the proper photography equipment to be acquired
before the flight. The focal length, flight lines and desired overlap, in combination with the size
of the area and chosen detail in output, determine the altitude of the mission and number of
images that are required to produce imagery suitable for the desired scale (Paine and Kiser
2012). Typical overlap in UAS mission plans is 80% while manned aerial vehicles require 60%
overlap for forward lap in the flight line and 30% on the side lap of each series of flight line
photographs. For 2D maps the flight path looks similar to a lawn mowing pattern, and for 3D
maps the flight path initially flies the same pattern as a 2D map; however, it then flips the pattern
90 degrees to cross over the original pattern so that a cross-hatch pattern is created. The 3D map
allows for digital elevation models to be created in programs such as Pix4D. The variables set
17

forth during the mission planning process need to be maintained during the flight to ensure
accurate results (Ahmad et al. 2013).
Mission planning differs between UAS and traditional aircraft. Mission planning
software that can be downloaded on iPhone and android phones can provide the UAS operator
with preliminary aerial imagery to use as a basemap to select the area that will be flown. These
missions can be created before the operator arrives on site or directly once a visual ground
assessment has occurred. The operator can then input the desired flight parameters such as
altitude and overlap percentage, as well as the camera details such as camera angle. The software
will create a flight path that covers the selected area, with precise points at which images will be
taken in order to gain the desired outputs. The mission is then uploaded to the UAS via a Wifi
link from the phone to the UAS and the mission begins when the operator starts the mission
within the application. The progress of the mission can be monitored on the computer in real-
time as the UAS completes the mission (Berteska and Ruzgiene 2013; Gademer et al. 2009).
Unlike UASs, planes and helicopters need to go through safety checks, repeated fuel up,
and can only take off once clearance is granted. These aircraft are limited by the available
airports in the area, which can be several miles from the study area, and could mean a
considerable time lag before images are captured. If refueling is needed to complete the mission
costs could continue to increase. On the other hand, most kinds of UASs can be taken to the site,
and within minutes, be launched and begin imagery collection. All missions will require ground
control points (GCPs), and these will need to be determined. GCPs can be temporary markers or
existing features on the ground that can be seen within the aerial photograph. The purpose of
GCPs is to provide locations on the image that can be precisely identified on the ground
(Campbell and Wynne 2011). The coordinates of the locations can be obtained through the use of
18

Global Positioning System (GPS) receivers in the field before, during, or after the mission. The
coordinates of the GCPs are used during post-processing to georeference the images to the
Earth’s surface. However, algorithms are being developed that might render GCPs unnecessary
in the future (Xiang and Tain 2011).
As with most methods, there are advantages and disadvantages to the utilization of each.
The light weight makes the UAS manageable to transport and assemble, but this light weight will
also make them more vulnerable to winds during flight that would not affect manned aerial
vehicles. Changes in wind, either direction or speed, can cause the UAS to pivot on its pitch, roll,
or yaw axis, all which can change the angle of the camera (Watkins et al. 2006). While wind
altered photos can be corrected it is best to plan for ideal weather conditions to save time in post
processing.
This last section has provided insight into the methods for preparing, obtaining, and
processing aerial imagery for mapping and monitoring purposes. Information on image overlap,
the use of mission planning software and GCP ensure proper coverage of the site. This
information served as a reference in the preparation for the data collection methods used for this
thesis.

19

Chapter 3: Methodology
This chapter describes the methods used in this thesis project. The case study locations are
introduced (Section 3.1), followed by a description of the equipment used to complete the study
(Section 3.2). The final section (Section 3.3) discusses data acquisition by UAS, field data
acquisition and post-processing of data for analysis.
3.1 Case Studies
The flights for the case studies occurred in May, June and July of 2018. The case studies
for this thesis consisted of 15 mapped NWI wetlands in Oklahoma. Four were located near
Meeker, OK, one was found in Pawnee, OK and ten were located near Roosevelt, OK. Meeker,
OK is located within the Northern Cross Timbers sub-region of the Cross Timbers ecoregion,
Pawnee, OK is located in the Cross Timbers Transition sub-region of the Great Plains ecoregion
and Roosevelt, OK is located in the Red Prairie sub-region of the Central Great Plains ecoregion
(Woods et al. 2005). The Northern Cross Timbers sub-region is characterized by mosaics of oak
savanna, scrubby oak forest, eastern red cedar (Juniperus virginiana), and tall grass prairie.
Areas within the ecoregion are used primarily for livestock farming, with cropland being less
extensive than in other ecoregions found in Oklahoma (Woods et al. 2005). The four case study
sites found near Meeker, OK are used in a seasonal rotation for grazing cattle. The Cross
Timbers Transition sub-region is characterized by rough plains that are covered by prairie
grasses and eastern red cedar, scattered oaks, and elms. Terrain and vegetation are transitional
between the less rugged grass-covered ecoregions to the west and the Northern Cross Timbers
sub-region to the east. Rangeland and livestock production is the primary land use (Woods et al.
2005). The Red-Prairie sub-region is characterized by mostly mesquite-buffalograss (Buchloe
dacryloides Nutt. Engelm) communities with gypsum outcrops. Wheat (Triticum sp.) is the main
20

crop, but unfavorable lands are maintained as rangelands. Like the sites in Meeker, OK, the
Pawnee and Roosevelt case study areas are used for cattle grazing with seasonal rotations.
3.2 Equipment
A DJI Phantom 3 Professional (DJI, Shenzhen, China) was used for imagery acquisition
on all case studies. The device is a commercial “all-in-one solution” quadcopter. In addition to
the aircraft itself, it consists of a built-in camera with a three-axis gimbal, remote control, a
mobile application for the aircraft and camera control, and information about the state of the
aircraft. The camera sensor is an RGB Sony Exmor 1/2.3” CMOS with lens FOV of 94◦, 20 mm
focal length, f/2.8 focal ratio, and focus to infinite. The image resolution of the camera is 4,000 ×
3,000 pixels (Table 1).
3.2.1 Systems and Software
The Phantom 3 Pro UAS (Figure 3) (DJI, Shenzhen, China) was chosen for this thesis
project due to accessibility, affordability, and compatibility with the mission planning software
(Pix4Dcapture). As seen in Table 1, the UAS with all accessories weighs approximately 1,280 g
and has a diagonal length of 350 mm. The UAS can have a flight time of 23 minutes (mins) with
a pristine lithium battery that is fully charged; however, this can be limited by subpar batteries or
adverse weather conditions. The UAS remote control unit operates between 2.400 and 2.483
GHz with a range of 3.1 mi (5 km) in unobstructed areas that are free of interference.

21

Table 1. List of field equipment and specifications
System Specifications
Aircraft - Phantom 3 Pro by DJI
Weight (Battery & Propellers Included) 1280 g
Diagonal Size (Propellers Excluded) 350 mm
Max Ascent Speed 5 m/s
Max Descent Speed 3 m/s
Max Speed 16 m/s (ATTI mode)
Max Tilt Angle 35°
Max Angular Speed 150°/s
Max Service Ceiling Above Sea Level 19685 feet (6000 m)
Max Flight Time Approx. 23 mins
Operating Temperature Range 32° to 104°F (0° to 40°C)
Satellite Positioning Systems GPS/GLONASS
Hover Accuracy Range Vertical:
±0.1 m (with Vision Positioning)
±0.5 m (with GPS Positioning)
Horizontal:
±0.3 m (with Vision Positioning)
±1.5 m (with GPS Positioning)
Gimbal
Stabilization 3-axis (pitch, roll, yaw)
Controllable Range Pitch: -90° to +30°
Max Controllable Angular Speed Pitch: 90°/s
Angular Control Accuracy ±0.02°
Remoter Control
Operating Frequency 2.400 - 2.483 GHz
Max Transmission Distance FCC Compliant: 3.1 mi (5 km)
CE Compliant: 2.2 mi (3.5 km)
(Unobstructed, free of interference)
Operating Temperature Range 32° to 104°F (0° to 40°C)
Battery 6000 mAh LiPo 2S
Transmitter Power (EIRP)
• FCC: 20 dBm
• CE: 16 dBm
MIC: 16 dBm
Operating Current/Voltage
• 1.2 A@7.4 V
Video Output Port USB
Mobile Device Holder Apple iPhone 7s
Camera
Sensor 1/2.3” CMOS
Effective pixels: 12.4 M (total pixels: 12.76 M)
Lens FOV 94° 20 mm (35 mm format equivalent) f/2.8 focus at ∞
ISO Range 100-1600 (photo)
Electronic Shutter Speed
• 8 - 1/8000 s
Image Size

• 4000×3000
Still Photography Modes
• Single Shot
• Burst Shooting: 3/5/7 frames
22

• Auto Exposure Bracketing (AEB): 3/5 bracketed frames at
0.7 EV Bias
Timelapse
Max Video Bitrate 60 Mbps
Supported File Systems
• FAT32 (≤32 GB); exFAT (>32 GB)
Photo JPEG, DNG (RAW)
Video MP4, MOV (MPEG-4 AVC/H.264)
Supported SD Cards Micro SD
Class 10 or UHS-1 rating required
Operating Temperature Range 32° to 104°F (0° to 40°C)
Other equipment

GPS Receiver Trimble Geo &X Handheld Model 88161
Trimble TerraSync 5.9 Software
Computers, Systems, and other software Dell HP Precision 7720
Esri ArcGIS 10.5 software
Pix4Dcapture mapping services
Pix4D Desktop 4.2.27 software
USFWS wetland Mapper
SanDisk Micro USB 32GB

Figure 3. Phantom 3 Pro UAS by DJI
(Source: https://www.dji.com/phantom-3-pro)
Images were collected via a SanDisk micro USB 32 GB. The Gimbal Roll Arm Motor
(Figure 4) by DJI is a 3-axis gimbal camera mount that secures the camera to the Phantom 3 Pro
UAS, providing stabilization for the camera should the UAS experience movements in “pitch”,
moving front to back; “rolling”, moving side to side; or “yaw”, moving left to right (Figure 5).
Without stabilization, such movements on a camera would cause the camera to tilt, roll, and pan.
23

Figure 4. Pro gimbal camera roll arm and camera
(Source: http://ww.dji.com)

Figure 5. Example of pitch, roll and yaw on UAS.
(Source: https://developer.dji.com/mobile-
sdk/documentation/introduction/flightController_concepts.html)

24

The Pix4dcapture ISO phone application supports 18 UAS models, from DJI and Parrot,
including the Phantom 3 Pro. This makes Pix4Dcapture ideal for mapping with no additional
equipment requirements (Draeyer and Strecha 2014). This application has specific programing
for the as-is Phantom 3 Pro by DJI, which allowed easy interface between the pilot and the
program with no adjustments needed. By using Pix4Dcapture, it allows the UAS pilot to plan
the mission (Figure 6), wireless upload it to the UAS, monitor the progress of the UAS during
flight, and land the UAS at the GPS point where the mission began. The pilot selects the total
area, drawing a square or polygon, to be flown based on Google Earth imagery or Google Street
Map. The pilot can select altitude, percent forward overlap and lateral overlap in order to gain
the desired resolution or adjust flight times. Higher altitudes and decrease percent overlap will
decrease flight time however resolution will also decrease.

Figure 6. Example of Pix4Dcapture mission
(Source: http://www.aerialpicture.co.za/the-pix4d-mapper-capture-app/)

25

3.2.2 Trimble GPS Receiver
A Trimble Geo 7X Series GPS receiver was used to collect the coordinates of the
sampled locations for each of the wetlands and ground control points on the aerial imagery. The
Geo 7X unit is a handheld unit that provides centimeter accuracy. While higher levels of
accuracy are considered more ideal, wetlands are a dynamic feature with fluctuations in growth
based on seasonal or yearly weather variability. Therefore, this level of accuracy was deemed
adequate for the purposes of this thesis. Trimble’s TerraSync 5.9 software installed on the GPS
receiver allows the unit to communicate with satellites and record positional data. The interface
allows the user to collect point, line, or area data, add attributes to the data, and store it in files
for download.
3.2.3 Software systems
NWI wetland mapper provided by the USFWS was reviewed to record if the wetland was
mapped and what year the wetland was most recently mapped, the wetland mapper was last
modified on May 1, 2018. It is important to note that this mapper is used to produce
reconnaissance level information on the location and size the features associated with the mapper
(USFWS 2018b)
Trimble’s GPS Pathfinder Office version 5.60 is a software platform installed on the Dell
HP Precision 7720 for use with the Trimble GPS receiver. The software allows the user to
upload data from TerraSync to the computer. GPS Pathfinder Office is used to perform post-data
correction by acquiring positional data from local base stations and improving the positional
accuracy of data collected with the GPS receiver. GPS Pathfinder Office also transforms data
files into shapefiles for use in the ArcGIS 10.5 software (Trimble Navigation Limited 2018).
Esri’s ArcGIS 10.5 software was used to perform the data analysis for this study.
26

Pix4Dmapper 4.2 was used to stitch together all of the images that were collected during
the UAS flights. Pix4Dmapper can process images taken from any angle from aerial or
terrestrial, manned or unmanned platforms. During processing this system allows for automatic
processing templates, optimization of internal camera parameters, such as focal length, principal
point of autocollimation and lens distortions, to compensate automatically for change in
brightness, luminosity and color balancing, and assess the accuracy and quality of projects with a
quality report. For this thesis the main output that was assessed from this system was the
orthomosaic image. Orthomosaic generation is the automated process for orthorectifing the raw
imagery and mosaicing adjacent images into one single large image. This process generated a
georeferenced image and optionally a digital surface model in various formats. The orthomosaic
is a 2D map. Each point contains X, Y and color information. The orthomosaic has a uniform
scale and can be used for 2D measurements (distance, surface).
Google Earth is used as a base map in Pix4Dcapture and provides preliminary basemap
imagery for mission planning. Google Earth was used to analyze existing imagery for
comparison between Esri base maps and UAS captures images in this thesis project.
3.3 Data Acquisition
The data acquisition via UAS and ground truthing occurred from May to July 2018 in
locations near Meeker, Pawnee and Roosevelt, OK.
3.3.1 UAS Data Acquisition
Prior to flights, permission to access the land was obtained and NWI wetland mapper was
reviewed to determine if the existing wetlands to be flown were currently mapped. If the
wetlands were not mapped, they were eliminated from the sample set. To prepare the UAS for
data collection, Pix4DCapture was opened on the ISO 9 iPhone 7, where the camera and flight
27

parameters were entered in the application. The data for this study was collected at an altitude of
40 m to avoid collisions with vegetation, which varied in height at each case study site. Percent
overlap for acquisition was set at 80% forward overlap and 80% lateral overlap between images;
the average flight speed was 0.5 m/s. Data collection was RGB at a high spatial resolution (2
inches per pixel). While LiDAR data is more accurate than RGB, this study was limited by the
cost of a LiDAR camera. Publicly available LiDAR data at the time of this thesis provides a
resolution of 10 m for the state of Oklahoma. This resolution would not provide robust indicators
of smaller slope changes found in areas with less distinct elevation changes. Next, a rectangle
was drawn over a portion of the case study area, with some flights encompassing multiple
mapped wetlands. This rectangle produced a preview of the flight path. Then the flight path was
created and uploaded to the UAS. Finally, a home point for the UAS was logged within
Pix4Dcapture and the UAS was launched via the application. The UAS uses the home point for
automatic landing at the end of the mission, or if the mission needs to be paused for battery
change. The time each flight took was recorded as well as time of day and cloud cover.
3.3.2 GPS Data Acquisition and Wetland delineation
Wetland delineation data for this thesis was collected with the Trimble GPS receiver. A
formal wetlands delineation was completed as an onsite assessment of three wetland indicators
(Section 1.2) as defined by the USACE Wetlands Delineation Manual (USACE 2010): hydric
soil, hydrology and hydrophilic vegetation. Presence and/or absence of all three indicators was
captured via the USACE Wetland Delineation Manual and the Midwest Regional Supplement
(USACE 2010) to determine the extent of the wetland. A series of data points that are
representative of the site were chosen and paired with wetland areas and upland sites. The
28

wetland border was recorded via the Geo 7X Trimble unit. The time each delineation required
was recorded as well.
3.3.3 Esri Base Maps and Google Earth Imagery Acquisition
Esri base maps are available online (www.arcgis.com) and can be used seamlessly with
ArcGIS 10.5 desktop software. Google Earth imagery was acquired via conversion of GIS
shapefiles of wetlands to KMZ files and uploaded to Google Earth 7.1. Both imagery datasets
were used for comparison with UAS acquired imagery. These datasets were chosen due to the
common use of the systems within wetland determination reports and other remotely sensed
reconnaissance reports.
3.4 Post-Processing Data
3.4.1 UAS Post-Processing
As covered in Section 3.2.2, UAS data consists of several hundred images that need to be
georeferenced and stitched together in order to be useful. Pix4DMapper software allows users to
stitch and georeference aerial imagery into maps. The points were located and matched
throughout the uploaded imagery to triangulate their positions in a manner that minimized the
errors between the points. Since the imagery has GPS information in an Exchangeable Image
File Format file, georeferencing occurred during the reconstruction of the imagery. An output
resolution of 15 cm was chosen for the project at hand. While the flight height of 40 m could
have yielded a resolution of 2 cm, a 15 cm resolution was selected to assess the processing and
analysis time. A higher altitude (120 m) could have been flown to obtain 15 cm resolution
imagery.

29

3.4.2 GPS Post-Processing
Ground-truthed data collected with the Trimble GPS receiver can have errors due to
overhead vegetation that interferes with satellite signal, or proximity to water. Differential
correction was performed within Trimble’s GPS Pathfinder Office software to correct errors that
occurred in the field. The differential correction process uses positional data from base station
providers in the area that is downloaded by the software. The base station records its location at
fixed intervals and compares it to the position of the control location to calculate the positional
error for the reading at that specific time. The time of day that the data collected by the GPS
receiver is cross-checked with the positional error data, applying positional error corrections to
data that were collected at corresponding times of day. The data files are converted to shapefile
format in GPS Pathfinder Office after differential correction (Dustin 2015).
3.5 Data Analysis
Processed orthomosiacs and wetland delineation polygons were uploaded into ArcGIS
10.5 for analysis. While Google Earth imagery cannot be uploaded into ArcGIS, a comparison of
resolution, and capture dates at similar scales was made.
3.5.1 Visual Comparison of UAS, Esri, and Google Earth Imagery
The UAS imagery is in WGS 84 and is projected to NAD 1983 UTM Zone 14 N using
the Project tool in ArcGIS. The NAD 1983 was used due to Oklahoma being divided into two
state plane systems, which would have added unneeded complications. NAD 1983 UTM Zone
14 N encompasses most of the state of Oklahoma (Figure 7) the conversion of the data from the
geographic coordinate system to a projected coordinate system allows measurements to be
calculated on the data. The UAS, Esri, and Google Earth imagery of the study site was opened
and visually examined. Ocular inspections of the imagery sets were performed, making note of
30

differences that can be seen in each, such as differences in season, clarity and if the image
registered the wetland in its entirety. After they were processed, supervised and unsupervised
classification was used to determine the wetland’s current extent and the surrounding vegetation.
Ten training classes were used to determine wetland extent, and isoclustering unsupervised
classification was used to classify the surrounding vegetation into three distinct vegetation
classes (forest, grassland, and bare ground).

Figure 7. Oklahoma NAD 1983 UTM zones (Source:
https://www.nrcs.usda.gov/Internet/FSE_DOCUMENTS/nrcs142p2_001664.pdf

NWI wetland data was downloaded from the USFWS wetland mapper website for the
state of Oklahoma. The case study wetlands were then selected and extracted from the statewide
dataset. The Calculate Geometry tool in ArcGIS measured the total number of acres for the UAS
imagery, NWI imagery and the GPS collected data. The calculated acreages from the UAS and
31

NWI imagery were then compared to the measurements captured with the GPS receiver during
the wetland delineation. Since wetland delineation is the highest form of wetland assessment,
this will be considered the truest measurement for current conditions. In order to assess the
accuracy of the UAS method versus the ground truth delineation of the wetland 200 accuracy
assessment points were created for each site. After accuracy assessment points were created, they
were input into the Compute Confusion Matrix Tool.
3.5.4 Cost/Benefit Analysis
The potential cost savings, when comparing UAS to traditional surveying techniques,
was noted in Chapter 2 and in order to assess this the mobilization time, flight times, and
processing times were recorded for all sites, as well as the time it took to perform a wetland
delineation on each site. The hourly rates of an UAS pilot and a certified wetland delineator
were captured and used with a 3.0 multiplier. These rates are compared to determine the
financial differences of the UAS to traditional wetland delineation. The total hours spent
collecting and processing data collected with the UAS and GPS receiver were tallied in order to
determine the total labor cost for each method of data acquisition. These estimates were
compared to estimate the potential labor savings benefits of the UAS. A SWOT (strength,
weakness, opportunity, and threat) analysis was also performed using the results of the financial
and labor benefits to evaluate the efficacy of using UAS technology for the purposes of wetland
determination. Information discussed in the literature review of Chapter 2 was also considered in
the SWOT analysis.

32

Chapter 4: Results
This chapter details the results of the methods documented in Chapter 3 to assess the
differences between UAS mapped wetlands and traditionally used data. As outlined earlier the
goals of this study were to: (1) improve upon current determination datasets; (2) determine the
accuracy of the imagery when compared to mapped NWI wetlands and wetland delineations; and
(3) determine the cost effectiveness of UAS when compared to traditional methods.
4.1 Accuracy of UAS and Existing data
The UAS-collected imagery was visually compared to Google Earth and ESRI imagery
available through ArcGIS Online in ArcGIS 10.5. The dates and resolution of the various
imagery data sets were first compared for relevance (Tables 2 - 4). Zooming in on the imagery
increases the amount of detail that can be seen. The resolution of the imagery determines the
amount of detail that can be observed before the imagery appears blurred or pixelated.
Table 2. Capture dates for imagery data in Meeker, OK case studies.
Imagery Source Capture Date Resolution
UAS 5/27/2018 5 cm
Google Earth 2/25/2014 15 cm
Esri World Imagery
Basemap
4/21/2016 50 cm

Table 3. Capture dates for imagery data in Pawnee, OK case study.
Imagery Source Capture Date Resolution
UAS 1/15/2018 and 6/10/2018 5 cm
Google Earth 2/25/2014 15 cm
Esri World Imagery
Basemap
4/2/2016 50 cm

Table 4. Capture dates for imagery data in Roosevelt, OK case studies.
Imagery Source Capture Date Resolution
UAS 7/1/2018 5 cm
Google Earth 10/14/2017 15 cm
Esri World Imagery
Basemap
7/10/2016 50 cm
33

Fifteen UAS missions were flown around NWI mapped wetlands (Figures 10 - 23).
Because wetlands are most accurately delineated during the growing season, flying during the
times when wetland delineations could take place concurrently seemed ideal. A full wetland
delineation was completed on each site in accordance with the USACE Wetland Delineation
Manual (USACE 1987) and the Midwest Regional Supplement (USACE 2010).
Weather plays a large role in current wetland size and status. Figures 8 and 9 show the
yearly variability in Oklahoma’s precipitation and temperature, due to this, historic weather
conditions were considered for each year that imagery was captured. Weather has an
accumulative effect on wetlands; therefore, the three months preceding the capture dates were
considered to see if the months had high or low temperatures and rainfall. Table 5 details the
weather conditions present in Meeker, OK during 2014, 2016 and 2018. Likewise, Tables 6 and
7 detail the weather conditions present for the capture dates as shown in previous tables for
Pawnee and Roosevelt, OK.

Figure 8. Oklahoma annual precipitation from 1895-2017
Source : http://climate.ok.gov/
34

Figure 9. Oklahoma annual temperature from 1895-2017
Source: http://climate.ok.gov/

Table 5. The average maximum, mean, and minimum temperatures and rainfall totals for
Meeker, OK three months prior to imagery capture dates.
Google Maps December 2013 January 2014 February 2014
Max temp (F) 62 60 62
Mean temp (F) 40 41 41
Min temp (F) 24 17 18
Precipitation (in.) 0.47 0.01 0.11

Esri Base Map February 2016 March 2016 April 2016
Max temp (F) 70 70 75
Mean temp (F) 49 56 62
Min temp (F) 36 42 51
Precipitation (in.) 1.68 0.91 4.48

UAS Mapping April 2018 May 2018 June 2018
Max temp (F) 72 80 83
Mean temp (F) 54 74 77
Min temp (F) 34 64 70
Precipitation (in.) 2.11 4.14 3.99

35

Table 6. The average maximum, mean, and minimum temperatures and rainfall totals for
Pawnee, OK three months prior to imagery capture dates.
Google Maps December 2013 January 2014 February 2014
Max temp (F) 60 54 54
Mean temp (F) 35 36 36
Min temp (F) 14 8 13
Precipitation (in.) 0.66 0.05 0.48

Esri Base Map February 2016 March 2016 April 2016
Max temp (F) 65 74 73
Mean temp (F) 47 55 61
Min temp (F) 34 40 49
Precipitation (in.) 0.26 2.52 4.61

UAS Mapping April 2018 May 2018 June 2018
Max temp (F) 75 82 84
Mean temp (F) 54 75 79
Min temp (F) 35 64 70
Precipitation (in.) 2.05 8.87 4.83

Table 7. The average maximum, mean, and minimum temperatures and rainfall totals for
Roosevelt, OK three months prior to imagery capture dates.
Google Maps April 2016 May 2016 June 2016
Max temp (F) 75 80 92
Mean temp (F) 63 69 80
Min temp (F) 50 58 69
Precipitation (in.) 6.06 4.82 2.61

Esri Base Map August 2017 September 2017 October 2017
Max temp (F) 89 86 77
Mean temp (F) 79 75 64
Min temp (F) 70 63 50
Precipitation (in.) 9.04 6.66 1.69

UAS Mapping April 2018 May 2018 June 2018
Max temp (F) 73 89 96
Mean temp (F) 57 77 84
Min temp (F) 41 64 71
Precipitation (in.) 1.44 2.80 1.46

36

The UAS imagery was collected at an altitude of 40 m during May, June and July of
2018, with one additional flight of the Pawnee case study occurring in January 2018. The UAS
imagery improved resolution and increased clarity. The Google Earth imagery, captured in
February 2014, is relevant to the extent that it shows the wetlands; however, there is a four-year
gap between the time these images were captured and the time at which this thesis was written.
The Google Earth imagery was captured during the winter months, which does not allow the user
to assess any of the vegetation patterns that exist in the growing season. The Esri base maps had
the lowest resolution of all the imagery; however, this imagery was captured during the growing
season. Each NWI wetland was plotted on the Esri base map and Table 8 shows the NWI
acreages and capture dates for each of the fifteen case studies. The USFWS Wetland Mapper was
last updated May 1, 2018.
Table 8. NWI mapped wetlands attributes
Case Study
ID
Wetland Type Classification* NWI Acreages Date of
mapping
1 Freshwater pond PubHh 1.32 03/80
2 Freshwater pond PubHh 0.44 03/80
3 Freshwater Pond PubHh 0.20 03/80
4 Freshwater Pond PubHh 0.94 03/80
5 Freshwater Pond PubHh 0.33 03/81
6 Freshwater pond PubHh 0.63 03/83
7 Freshwater pond PubHh 1.67 03/83
8 Freshwater Pond PubHh 1.39 03/83
9 Freshwater Pond PubHx 0.51 03/83
10 Freshwater Pond PubFh 0.73 03/83
11 Freshwater pond PubHh 0.35 03/83
12 Freshwater pond PubHh 0.12 03/83
13 Freshwater Pond PubFh 0.11 03/83
14 Freshwater Pond PubHx 0.67 03/83
15 Freshwater Pond PubFx 0.5 03/83
* See next page for classification description.
The NWI data that has been collected shows that all of the wetlands flown where
consistent in type and classification. There is a range of 35-37 years between the mapping of
37

these wetlands and the preparation of this thesis. The case study NWI wetlands range from 0.33
acres to 1.67 acres in size (Table 8). The Cowardian classification (USEPA 2002) code used by
the USFWS and reported in column 3 of Table 8 states that all wetlands for this thesis include
the following traits:
• P - PALUSTRINE: The Palustrine System includes all nontidal wetlands
dominated by trees, shrubs, persistent emergents, emergent mosses or lichens, and
all such wetlands that occur in tidal areas where salinity due to ocean-derived
salts is below 0.5 ppt. It also includes wetlands lacking such vegetation, but with
all of the following four characteristics: (a) area less than 8 ha (20 acres); (b)
active wave-formed or bedrock shoreline features lacking; (c) water depth in the
deepest part of basin less than 2.5 m (8.2 ft) at low water; and (d) salinity due to
ocean-derived salts less than 0.5 ppt.
• UB - UNCONSOLIDATED BOTTOM: Includes all wetlands and deep-water
habitats with at least 25% cover of particles smaller than stones (less than 6-7
cm), and a vegetative cover less than 30%.
• H - Water Regime Permanently Flooded: Water covers the substrate throughout
the year in all years.
• h - SPECIAL MODIFIER Diked/Impounded: These wetlands have been created or
modified by a man-made barrier or dam that obstructs the inflow or outflow of
water.
• F - Water Regime Semipermanently Flooded: Surface water persists throughout
the growing season in most years. When surface water is absent, the water table is
usually at or very near the land surface.
38

• x - SPECIAL MODIFIER Excavated: This modifier is used to identify wetland
basins or channels that were excavated by humans.
4.2 Comparison of NWI, Delineation, and UAS Collection Methods
From the flight, delineation and NWI data, there are noticeable differences in the size
(Table 9) and placement of the wetlands (Figures 8 – 22). There was greater overlap between the
NWI and the delineation data compared to the UAS data; however, between the three data sets
the highest percent overlap was found between the delineation data and the UAS collected
imagery. Overall the UAS mapped wetlands offer more conservative measurements of area than
the full delineation.
Table 9. Acreages calculated from NWI, Delineation, and UAS collection methods.

The following case studies document the time, weather conditions and vegetation present
at the time of data collection. The wetlands were split into 3 categories: wetlands without
obstruction; wetlands with canopy cover/blue green algae; and wetland with exposed bed. From
these three categories 98.3%, 93.4% and 85% accuracy, respectively, was calculated.
Case Study ID NWI Acreages Delineation UAS
1 1.32 1.47 1.29
2 0.44 0.67 0.49
3 0.20 0.35 0.34
4 0.94 2.34 1.80
5 0.33 0.87 0.62
6 0.63 0.81 0.73
7 1.67 0.87 0.72
8 1.39 1.68 1.24
9 0.51 0.86 0.75
10 0.73 0.65 0.65
11 0.35 0.67 0.54
12 0.12 0.23 0.1
13 0.11 0.27 0.2
14 0.67 0.59 0.29
15 0.5 0.06 0.005
39

4.2.1 Case Study #1
Case Study #1 was flown on May 28, 2018 from 5:00 to 5:25 p.m. with a total flight time
of 25 minutes and 36 seconds; the flight collected data for Case Study #2 concurrently. Weather
conditions were clear skies with winds of 8-12 miles per hour. Delineation of the wetland took
place from 5:30 to 9:10 p.m. (i.e. 3 hours and 40 minutes). The wetland was surrounded by
Blackjack oak (Quercus marilandica), Post oak (Quercus stellate), and green briar (Smilax sp.).
Prominent wetland species that were present included Rufous bulrush (Scirpus pendulus), Sago
Pondweed (Stuckenia pectinate) and blue green algae (Cyanobacteria sp.). Figure 10a displays
the Esri basemap, captured in April 2016, and the historic NWI mapped wetland (1.32 ac),
captured in March 1980. There is a 36-year time lapse between these two data sets. With the
coarse imagery provided at the 1:1,500 scale, the NWI mapped area does not fully represent the
2016 extent of the wetland; however, the placement is centered over the present wetland. In
Figure 10b the resolution is increased but the discrepancy between the NWI data and visual
estimation of the wetland location is still present. Figure 10c shows the delineated area (1.47 ac)
and historic wetland overlap, which indicates a 7.6 % increase in wetland size from NWI. Figure
10d shows the wetland area determined by supervised classification via UAS imagery (1.29 ac)
and historic wetland overlap. When comparing overall acreage of the UAS versus the historic
wetland, there is a 2.27% decrease in wetland size. The UAS collected data provides a more
conservative estimate than a full delineation, the delineation shows a small increase in wetland
size when compared to the historic data captured 37 years ago, while the UAS method indicates
a small decrease. Finally, Figure 10e shows the overlap between the wetland delineation and the
UAS mapped wetland. Ninety-nine percent of the UAS mapped wetland area is found within the
delineated area, while the UAS mapped area covers approximately 87.7% of the delineated area.
40

Figure 10. Case Study # 1 in Meeker, OK: (a) Esri base map with NWI mapped wetland; (b) UAS imagery with NWI mapped
wetland; (c) Wetland Delineation results, NWI mapped wetland and overlap between the two; (d) UAS mapped wetland results, NWI
mapped wetland and overlap between the two; and (e) Overlap between Delineation and UAS mapped wetland results.
41

4.2.2 Case Study #2
Case Study #2 was flown on May 28, 2018 from 5:00 to 5:25 p.m. with a total flight time of 25
minutes and 36 seconds; the flight collected data for Case Study #1 concurrently. Weather
conditions were clear skies with winds of 8-12 miles per hour. Delineation of the wetland took
place from 9 to 10:30 a.m. on May 29, 2018 (i.e. one hour and 30 minutes). The wetland was
surrounded by Blackjack oak, eastern red cedar, and green briar. Wetland species that were
present included blue green algae, but disturbance was present around the shore and other
vegetation was removed. Figure 11a displays the Esri basemap, captured in April 2016, and the
historic NWI mapped wetland (0.44 acres), captured in March 1981. There is a 35-year time
lapse between these two data sets. Even with the coarse imagery provided at the 1:1,000 scale,
the NWI mapped area does not represent the 2016 extent of the wetland. In Figure 11b the
resolution is increased and the discrepancy between the NWI data and visual estimation of the
wetland location is still present. Figure 11c shows the total delineated area (0.67 acres) and
historic wetland overlap. When comparing overall acreage of historic wetland versus the
delineated area, the delineation method indicates a 52% increase in wetland size. Figure 11d
shows the wetland area determined by supervised classification via UAS imagery (0.49 acres)
and historic wetland overlap. When comparing overall acreage of the UAS versus the historic
wetland, there is a 11% increase in wetland size. While the UAS collected data provides a more
conservative estimate than a full delineation both show an increase in wetland size when
compared to the historic data captured 37 years ago. Finally, Figure 11e shows the overlap
between the wetland delineation and the UAS mapped wetland. Ninety-nine percent of the UAS
mapped wetland area is found within the delineated area, while the UAS mapped area covers
approximately 73.1% of the delineated area.
42

Figure 11. Case Study # 2 in Meeker, OK: (a) Esri base map with NWI mapped wetland; (b) UAS imagery with NWI mapped
wetland; (c) Wetland Delineation results, NWI mapped wetland and overlap between the two; (d) UAS mapped wetland results, NWI
mapped wetland and overlap between the two; and (e) Overlap between Delineation and UAS mapped wetland results.
43

4.2.3. Case Study #3
Case Study #3 was flown on May 28, 2018 from 2:30 to 2:58 p.m. with a total flight time of 28
minutes and 3 seconds; this flight also collected the data for Case study #4. Weather conditions
were clear skies with winds of 7-12 miles per hour. Delineation of the wetland occurred took
place from 11 to 11:30 a.m. (i.e. 30 minutes). The wetland was surrounded by pasture land,
switchgrass (Panicum virgatum), sideoats grama (Bouteloua curtipendula), hairy grama
(Bouteloua hirsute), and bermuda grass (Cynodon dactylon). No wetland species were present.
Figure 12a displays the Esri basemap, captured in April 2016, and the historic NWI mapped
wetland (0.20 acres), captured in March 1981. There is a 35-year time lapse between these two
data sets. Even with the coarse imagery provided at the 1:750 scale, the NWI mapped area does
not represent the 2016 extent of the wetland. In Figure 12b the resolution is increased but the
discrepancy between the NWI data and visual estimation of the wetland location is still present.
Figure 12c shows the total delineated area (0.35 acres) and historic wetland overlap. When
comparing overall acreage of historic wetland versus the delineated area, the delineation method
indicates a 75% increase in wetland size. Figure 12d shows the wetland area determined by
supervised classification via UAS imagery (0.34 acres) and historic wetland overlap. When
comparing overall acreage of the UAS versus the historic wetland, there is a 70% increase in
wetland size. While the UAS collected data provides a more conservative estimate than a full
delineation both show a marked increase in wetland size when compared to the historic data
captured 37 years ago. Finally, Figure 12e shows the overlap between the wetland delineation
and the UAS mapped wetland. Ninety-nine percent of the UAS mapped wetland area is found
within the delineated area, while the UAS mapped area covers approximately 97.1% of the
delineated area.
44

Figure 12. Case Study # 3 in Meeker, OK: (a) Esri base map with NWI mapped wetland; (b) UAS imagery with NWI mapped
wetland; (c) Wetland Delineation results, NWI mapped wetland and overlap between the two; (d) UAS mapped wetland results, NWI
mapped wetland and overlap between the two; and (e) Overlap between Delineation and UAS mapped wetland results.
45

4.2.4 Case Study #4
Case Study #4 was flown on June 6, 2018 from 5:26 to 5:36 p.m. with a total flight time
of 10 minutes and 10 seconds. Weather conditions were clear skies with winds of 5-12 miles per
hour. Delineation of the wetland took place from 2:45 to 5:10 p.m. (i.e. two hours and 25
minutes). The wetland was surrounded by Blackjack oak, eastern red cedar, and green briar.
Wetland species that were present included Rufous bulrush and Shoreline sedge. Figure 13a
displays the Esri basemap, captured in April 2016, and the historic NWI mapped wetland (0.94
acres), captured in March 1981. There is a 35-year time lapse between these two data sets. Even
with the coarse imagery provided at the 1:2,500 scale, the NWI mapped area does not represent
the 2016 extent of the wetland. In Figure 13b the resolution is increased but the discrepancy
between the NWI data and visual estimation of the wetland location is still present. Figure 13c
shows the total delineated area (2.34 acres) and historic wetland overlap. When comparing
overall acreage of historic wetland versus the delineated area, the delineation method indicates a
264% increase in wetland size. Figure 13d shows the wetland area determined by supervised
classification via UAS imagery (1.80 acres) and historic wetland overlap. When comparing
overall acreage of the UAS versus the historic wetland, there is a 189% increase in wetland size.
While the UAS collected data provides a more conservative estimate than a full delineation both
show a marked increase in wetland size when compared to the historic data captured 37 years
ago. Finally, Figure 13e shows the overlap between the wetland delineation and the UAS
mapped wetland. Ninety-nine percent of the UAS mapped wetland area is found within the
delineated area, while the UAS mapped area covers approximately 87.8% of the delineated area.
46

Figure 13. Case Study # 4 in Meeker, OK: (a) Esri base map with NWI mapped wetland; (b) UAS imagery with NWI mapped
wetland; (c) Wetland Delineation results, NWI mapped wetland and overlap between the two; (d) UAS mapped wetland results, NWI
mapped wetland and overlap between the two; and (e) Overlap between Delineation and UAS mapped wetland results.
47

4.2.5 Case Study #5
Case Study #5 was flown on June 6, 2018 from 5:26 to 5:36 p.m. with a total flight time
of 10 minutes and 10 seconds. Weather conditions were clear skies with winds of 5-12 miles per
hour. Delineation of the wetland took place from 2:45 to 5:10 p.m. (i.e two hours and 25
minutes). The wetland was surrounded by Blackjack oak, eastern red cedar, and green briar.
Wetland species that were present included Rufous bulrush, Shoreline sedge, and blue green
algae. Figure 14a displays the Esri basemap, captured in April 2016, and the historic NWI
mapped wetland (0.33 acres), captured in March 1981. There is a 35-year time lapse between
these two data sets. Even with the coarse imagery provided at the 1:1,500 scale, the NWI
mapped area does not represent the 2016 extent of the wetland. In Figure 14b the resolution is
increased but the discrepancy between the NWI data and visual estimation of the wetland
location is still present. Figure 14c shows the total delineated area (0.87 acres) and historic
wetland overlap. When comparing overall acreage of historic wetland versus the delineated area,
the delineation method indicates a 163% increase in wetland size. Figure 14d shows the wetland
area determined by supervised classification via UAS imagery (0.62 acres) and historic wetland
overlap. When comparing overall acreage of the UAS versus the historic wetland, there is an
87.9% increase in wetland size. While the UAS collected data provides a more conservative
estimate than a full delineation both show a marked increase in wetland size when compared to
the historic data captured 37 years ago. Finally, Figure 14e shows the overlap between the
wetland delineation and the UAS mapped wetland. Ninety-nine percent of the UAS mapped
wetland area is found within the delineated area, while the UAS mapped area covers
approximately 87.8% of the delineated area.
48

Figure 14. Case Study # 5 in Pawnee, OK: (a) Esri base map with NWI mapped wetland; (b) UAS imagery with NWI mapped
wetland; (c) Wetland Delineation results, NWI mapped wetland and overlap between the two; (d) UAS mapped wetland results, NWI
mapped wetland and overlap between the two; and (e) Overlap between Delineation and UAS mapped wetland results.
49

4.2.6. Case Study #6
Case Study #6 was flown on June 21, 2018 from 11:30 to 11:42 p.m. with a total flight
time of 12 minutes and 46 seconds. Weather conditions were overcast skies with winds of 5-12
miles per hour. Delineation of the wetland took place from 11 to 11:30 a.m. (i.e. 30 minutes).
The wetland was surrounded by pasture land with little bluestem (Schizachyrium scoparium),
Indian grass (Sorghastrum nutans), big bluestem (Andropogon gerardii), switchgrass (Panicum
virgatum), sideoats grama (Bouteloua curtipendula), hairy grama (Bouteloua hirsute), and blue
grama (Bouteloua gracilis) bermuda grass (Cynodon dactylon). Figure 15a displays the Esri
basemap, captured in July 2016, and the historic NWI mapped wetland (0.63 acres), captured in
March 1983. There is a 33-year time lapse between these two data sets. Even with the coarse
imagery provided at the 1:1,250 scale, the NWI mapped area does not represent the 2016 extent
of the wetland. In Figure 15b the resolution is increased via UAS imagery but the discrepancy
between the NWI data and visual estimation of the wetland location is still present. Figure 15c
shows the total delineated area (0.81 acres) and historic wetland overlap, with the delineation
method indicating a 28.6% increase in wetland size. Figure 15d shows the wetland area
determined by supervised classification via UAS imagery (0.73 acres) and historic wetland
overlap. When comparing overall acreage of the UAS versus the historic wetland, there is an
15.9% increase in wetland size. While the UAS collected data provides a more conservative
estimate than a full delineation both show an increase in wetland size when compared to the
historic data captured 35 years ago. Finally, Figure 15e shows the overlap between the wetland
delineation and the UAS mapped wetland. One hundred percent of the UAS mapped wetland
area is found within the delineated area, while the UAS mapped area covers approximately
90.1% of the delineated area
50

Figure 15. Case Study # 6 in Roosevelt, OK: (a) Esri base map with NWI mapped wetland; Figure (b) UAS imagery with NWI
mapped wetland; (c) Wetland Delineation results, NWI mapped wetland and overlap between the two; (d) UAS mapped wetland
results, NWI mapped wetland and overlap between the two; and (e) Overlap between Delineation and UAS mapped wetland results.
51

4.2.7 Case Study #7
Case Study #7 was flown on June 21, 2018 from 12:15 to 12:25 p.m. with a total flight time of
10 minutes and 13 seconds. Weather conditions were 20% cloud cover with winds of 8-15 miles
per hour. Delineation of the wetland took place from 12:30 to 1:27 p.m. (i.e. 57 minutes). The
wetland was surrounded by pasture land, with little bluestem, Indian grass, big bluestem,
switchgrass, sideoats grama, and bermuda grass. No wetland species were identified. Figure 16a
displays the Esri basemap, captured in July 2016, and the historic NWI mapped wetland (1.67
acres), captured in March 1983. There is a 33-year time lapse between these two data sets. Even
with the coarse imagery provided at the 1:1,750 scale, the NWI mapped area does not represent
the 2016 extent of the wetland. In Figure 16b the resolution is increased via UAS imagery but the
discrepancy between the NWI data and visual estimation of the wetland location is still present.
Figure 16c shows the total delineated area (0.87 acres) and historic wetland overlap. When
comparing overall acreage of historic wetland versus the delineated area, the delineation method
indicates a 47.9% decrease in wetland size. Figure 16d shows the wetland area determined by
supervised classification via UAS imagery (0.72 acres) and historic wetland overlap. When
comparing overall acreage of the UAS versus the historic wetland, there is an 56.9% decrease in
wetland size. While the UAS collected data provides a more conservative estimate than a full
delineation both show a marked decrease in wetland size when compared to the historic data
captured 35 years ago. Finally, Figure 16e shows the overlap between the wetland delineation
and the UAS mapped wetland. One hundred percent of the UAS mapped wetland area is found
within the delineated area, while the UAS mapped area covers approximately 82.8% of the
delineated area.
52

Figure 16. Case Study # 7 in Roosevelt, OK: (a) Esri base map with NWI mapped wetland; (b) UAS imagery with NWI mapped
wetland; (c) Wetland Delineation results, NWI mapped wetland and overlap between the two; (d) UAS mapped wetland results, NWI
mapped wetland and overlap between the two; and (e) Overlap between Delineation and UAS mapped wetland results.
53

4.2.8 Case Study #8
Case Study #8 was flown on July 1, 2018 from 11:30 to 12:22 p.m. with a total flight
time of 52 minutes; this flight also collected data for Case Studies #9 and #10. Weather
conditions were overcast skies with winds of ≤10 miles per hour. Delineation of the wetland took
place from 9:00 to 9:25 a.m. (i.e. 25 minutes). The wetland was surrounded by pasture land, with
little bluestem, Indian grass, big bluestem, switchgrass, sideoats grama, and bermuda grass. No
wetland species were identified. Figure 17a displays the Esri basemap, captured in July 2016,
and the historic NWI mapped wetland (1.39 acres), captured in March 1983. There is a 33-year
time lapse between these two data sets. Even with the coarse imagery provided at the 1:2,000
scale, the NWI mapped area does not represent the 2016 extent of the wetland. In Figure 17b the
resolution is increased via UAS imagery but the discrepancy between the NWI data and visual
estimation of the wetland location is still present. Figure 17c shows the total delineated area
(1.68 acres) and historic wetland overlap. When comparing overall acreage of historic wetland
versus the delineated area, the delineation method indicates a 20.9% increase in wetland size.
Figure 17d shows the wetland area determined by supervised classification via UAS imagery
(1.24 acres) and historic wetland overlap. When comparing overall acreage of the UAS versus
the historic wetland, there is an 10.8% decrease in wetland size. The UAS collected data
provides a more conservative estimate than a full delineation, with the delineation showing an
increase in wetland size and UAS mapped area showing a decrease in size when compared to the
historic data captured 35 years ago. Finally, Figure 17e shows the overlap between the wetland
delineation and the UAS mapped wetland. One hundred percent of the UAS mapped wetland
area is found within the delineated area, while the UAS mapped area covers approximately
80.6% of the delineated area.
54

Figure 17. Case Study # 8 in Roosevelt, OK: (a) Esri base map with NWI mapped wetland; (b) UAS imagery with NWI mapped
wetland; (c) Wetland Delineation results, NWI mapped wetland and overlap between the two; (d) UAS mapped wetland results, NWI
mapped wetland and overlap between the two; and (e) Overlap between Delineation and UAS mapped wetland results.
55

4.2.9 Case Study #9
Case Study #9 was flown on July 1, 2018 from 11:30 to 12:22 p.m. with a total flight
time of 52 minutes; this flight also collected data for Case Studies #8 and #10. Weather
conditions were overcast skies with winds of ≤ 10 miles per hour. Delineation of the wetland
took place from 9:30 to 10:15 a.m. (i.e. 45 minutes). The wetland was surrounded by pasture
land with species such as honey mesquite (Prosopis glandulose), little bluestem, Indian grass,
big bluestem, switchgrass, sideoats grama, and bermuda grass. Wetland species included
duckweed (Lemnoideae sp.). Figure 18a displays the Esri basemap, captured in July 2016, and
the historic NWI mapped wetland (0.51 acres), captured in March 1983. There is a 33-year time
lapse between these two data sets. Even with the coarse imagery provided at the 1:1,250 scale,
the NWI mapped area does not represent the 2016 extent of the wetland. In Figure 18b the
resolution is increased via UAS imagery but the discrepancy between the NWI data and visual
estimation of the wetland location is still present. Figure 18c shows the total delineated area
(0.86 acres) and historic wetland overlap. When comparing overall acreage of historic wetland
versus the delineated area, the delineation method indicates a 68.6% increase in wetland size.
Figure 18d shows the wetland area determined by supervised classification via UAS imagery
(0.75 acres) and historic wetland overlap. When comparing overall acreage of the UAS versus
the historic wetland, there is an 47.1% increase in wetland size. While the UAS collected data
provides a more conservative estimate than a full delineation both show a marked increase in
wetland size when compared to the historic data captured 35 years ago. Finally, Figure 18e
shows the overlap between the wetland delineation and the UAS mapped wetland. Ninety-six
percent of the UAS mapped wetland area is found within the delineated area, while the UAS
mapped area covers approximately 88.3% of the delineated area.
56

Figure 18. Case Study # 9 in Roosevelt, OK: (a) Esri base map with NWI mapped wetland; (b) UAS imagery with NWI mapped
wetland; (c) Wetland Delineation results, NWI mapped wetland and overlap between the two; (d) UAS mapped wetland results, NWI
mapped wetland and overlap between the two; and (e) Overlap between Delineation and UAS mapped wetland results.
57

4.2.10 Case Study #10
Case Study #10 was flown on July 1, 2018 from 11:30 to 12:22 p.m. with a total flight
time of 52 minutes; this flight also collected data for Case Studies #8 and #9. Weather conditions
were overcast skies with winds of ≤10 miles per hour. Delineation of the wetland took place
from 10:25 to 11:20 a.m. (i.e. 55 minutes). The wetland was surrounded by pasture land with
species such as honey mesquite, little bluestem, Indian grass, big bluestem, switchgrass, sideoats
grama, and bermuda grass. Wetland species included duckweed and blue green algae. Figure 19a
displays the Esri basemap, captured in July 2016, and the historic NWI mapped wetland (0.73
acres), captured in March 1983. There is a 33-year time lapse between these two data sets. Even
with the coarse imagery provided at the 1:1,000 scale, the NWI mapped area does not represent
the 2016 extent of the wetland. In Figure 19b the resolution is increased via UAS imagery but the
discrepancy between the NWI data and visual estimation of the wetland location is still present.
Figure 19c shows the total delineated area (0.65 acres) and historic wetland overlap. When
comparing overall acreage of historic wetland versus the delineated area, the delineation method
indicates a 11% decrease in wetland size. Figure 19d shows the wetland area determined by
supervised classification via UAS imagery (0.65 acres) and historic wetland overlap. When
comparing overall acreage of the UAS versus the historic wetland, there is an 11% decrease in
wetland size. While the UAS collected data provides the same estimation of area as a full
delineation both shows a decrease in wetland size when compared to the historic data captured
35 years ago. Finally, Figure 19e shows the overlap between the wetland delineation and the
UAS mapped wetland. One hundred percent of the UAS mapped wetland area is found within
the delineated area, while the UAS mapped area covers approximately 99.8% of the delineated
area.
58

Figure 19. Case Study # 10 in Roosevelt, OK: (a) Esri base map with NWI mapped wetland; (b) UAS imagery with NWI mapped
wetland; (c) Wetland Delineation results, NWI mapped wetland and overlap between the two; (d) UAS mapped wetland results, NWI
mapped wetland and overlap between the two; and (e) Overlap between Delineation and UAS mapped wetland results
59

4.2.11 Case Study #11
Case Study #11 was flown on July 1, 2018 from 12:45 to 1:48 p.m. with a total flight
time of 1 hour and 3 minutes; this flight collected data for Case Studies #11 to #15. Weather
conditions were overcast skies with winds of ≤10 miles per hour. Delineation of the wetland took
place from 2:00 to 2:33 p.m. (i.e. 33 minutes). The wetland was surrounded by pasture land with
species such as honey mesquite, little bluestem, Indian grass, big bluestem, switchgrass, sideoats
grama, and bermuda grass. Wetland species included duckweed and blue green algae. Figure 20a
displays the Esri basemap, captured in July 2016, and the historic NWI mapped wetland (0.35
acres), captured in March 1983. There is a 33-year time lapse between these two data sets. Even
with the coarse imagery provided at the 1:1,500 scale, the NWI mapped area does not represent
the 2016 extent of the wetland. In Figure 20b the resolution is increased via UAS imagery but the
discrepancy between the NWI data and visual estimation of the wetland location is still present.
Figure 20c shows the total delineated area (0.67 acres) and historic wetland overlap. When
comparing overall acreage of historic wetland versus the delineated area, the delineation method
indicates a 97.4% increase in wetland size. Figure 20d shows the wetland area determined by
supervised classification via UAS imagery (0.54 acres) and historic wetland overlap. When
comparing overall acreage of the UAS versus the historic wetland, there is an 54.3% increase in
wetland size. While the UAS collected data provides a more conservative estimate than a full
delineation both show a marked increase in wetland size when compared to the historic data
captured 35 years ago. Finally, Figure 20e shows the overlap between the wetland delineation
and the UAS mapped wetland. One hundred percent of the UAS mapped wetland area is found
within the delineated area, while the UAS mapped area covers approximately 80.6% of the
delineated area.
60

Figure 20. Case Study # 11 in Roosevelt, OK: (a) Esri base map with NWI mapped wetland; (b) UAS imagery with NWI mapped
wetland; (c) Wetland Delineation results, NWI mapped wetland and overlap between the two; (d). UAS mapped wetland results, NWI
mapped wetland and overlap between the two; and (e) Overlap between Delineation and UAS mapped wetland results.
61

4.2.12. Case Study #12
Case Study #12 was flown on July 1, 2018 from 12:45 to 1:48 p.m. with a total flight
time of 1 hour and 3 minutes; this flight collected data for Case Studies #11 to #15. Weather
conditions were overcast skies with winds of ≤10 miles per hour. Delineation of the wetland took
place from 2:35 to 3:00 p.m. (i.e. 25 minutes). The wetland was surrounded by pasture land with
species such as honey mesquite, little bluestem, Indian grass, big bluestem, switchgrass, sideoats
grama, and bermuda grass. Wetland species included duckweed and sago pondweed. Figure 21a
displays the Esri basemap, captured in July 2016, and the historic NWI mapped wetland (0.12
acres), captured in March 1983. There is a 33-year time lapse between these two data sets. Even
with the coarse imagery provided at the 1:750 scale, the NWI mapped area does not represent the
2016 extent of the wetland. In Figure 21b the resolution is increased via UAS imagery but the
discrepancy between the NWI data and visual estimation of the wetland location is still present.
Figure 21c shows the total delineated area (0.23 acres) and historic wetland overlap, with the
delineation method indicating a 91.7% increase in wetland size. Figure 21d shows the wetland
area determined by supervised classification via UAS imagery (0.10 acres) and historic wetland
overlap. When comparing overall acreage of the UAS versus the historic wetland, there is an
16.7% decrease in wetland size. The UAS collected data provides a more conservative estimate
than a full delineation, with the full delineation indicating an increase in wetland size and the
UAS mapped wetland indicating a decrease in wetland size when compared to the historic data
captured 35 years ago. Finally, Figure 21e shows the overlap between the wetland delineation
and the UAS mapped wetland. Ninety-five percent of the UAS mapped wetland area is found
within the delineated area, while the UAS mapped area covers approximately 43.5% of the
delineated area.
62

Figure 21. Case Study # 12 in Roosevelt, OK: (a) Esri base map with NWI mapped wetland; (b) UAS imagery with NWI mapped
wetland; (c) Wetland Delineation results, NWI mapped wetland and overlap between the two; (d) UAS mapped wetland results, NWI
mapped wetland and overlap between the two; and (e) Overlap between Delineation and UAS mapped wetland results.
63

4.2.13. Case Study #13
Case Study #13 was flown on July 1, 2018 from 12:45 p.m. to 1:48 p.m. with a total
flight time of 1 hour and 3 minutes; this flight collected data for Case Studies #11 to #15.
Weather conditions were overcast skies with winds of ≤10 miles per hour. Delineation of the
wetland took place from 3:10 to 3:52 p.m. (i.e. 42 minutes). The wetland was surrounded by
pasture land with species such as honey mesquite, little bluestem, with Indian grass, big
bluestem, switchgrass, sideoats grama, and bermuda grass. No wetland species were identified.
Figure 22a displays the Esri basemap, captured in July 2016, and the historic NWI mapped
wetland (0.11 acres), captured in March 1983. There is a 33-year time lapse between these two
data sets. Even with the coarse imagery provided at the 1:750 scale, the NWI mapped area does
not represent the 2016 extent of the wetland. In Figure 22b the resolution is increased but the
discrepancy between the NWI data and visual estimation of the wetland location is still present.
Figure 22c shows the total delineated area (0.27 acres) and historic wetland overlap. When
comparing overall acreage of historic wetland versus the delineated area, the delineation method
indicates a 145.6% increase in wetland size. Figure 22d shows the wetland area determined by
supervised classification via UAS imagery (0.20 acres) and historic wetland overlap. When
comparing overall acreage of the UAS versus the historic wetland, there is an 81.8% increase in
wetland size. While the UAS collected data provides a more conservative estimate than a full
delineation both show a marked increase in wetland size when compared to the historic data
captured 35 years ago. Finally, Figure 22e shows the overlap between the wetland delineation
and the UAS mapped wetland. One hundred percent of the UAS mapped wetland area is found
within the delineated area, while the UAS mapped area covers approximately 74.1% of the
delineated area.
64

Figure 22. Case Study # 13 in Roosevelt, OK: (a) Esri base map with NWI mapped wetland; (b) UAS imagery with NWI mapped
wetland; (c) Wetland Delineation results, NWI mapped wetland and overlap between the two; (d) UAS mapped wetland results, NWI
mapped wetland and overlap between the two; and (e) Overlap between Delineation and UAS mapped wetland results
65

4.2.14. Case Study #14
Case Study #14 was flown on July 1, 2018 from 12:45 to 1:48 p.m. with a total flight
time of 1 hour and 3 minutes; this flight collected data for Case Studies #11 to #15. Weather
conditions were overcast skies with winds of ≤10 miles per hour. Delineation of the wetland took
place from 4:00 to 4:30 p.m. (i.e. 30 minutes). The wetland was surrounded by pasture land with
species such as honey mesquite, little bluestem, Indian grass, big bluestem, switchgrass, sideoats
grama, and bermuda grass. The only wetlands species was duckweed. Figure 23a displays the
Esri basemap, captured in April 2016, and the historic NWI mapped wetland (0.67 acres),
captured in March 1983. There is a 33-year time lapse between these two data sets. Even with
the coarse imagery provided at the 1:1,250 scale, the NWI mapped area does not represent the
2016 extent of the wetland. In Figure 23b the resolution is increased via UAS imagery but the
discrepancy between the NWI data and visual estimation of the wetland location is still present.
Figure 23c shows the total delineated area (0.59 acres) and historic wetland overlap. When
comparing overall acreage of historic wetland versus the delineated area, the delineation method
indicates an 88% decrease in wetland size. Figure 23d shows the wetland area determined by
supervised classification via UAS imagery (0.29 acres) and historic wetland overlap. When
comparing overall acreage of the UAS versus the historic wetland versus the UAS, there is an
43.3% decrease in wetland size. While the UAS collected data provides a more conservative
estimate than a full delineation both show a marked decrease in wetland size when compared to
the historic data captured 35 years ago. Finally, Figure 23e shows the overlap between the
wetland delineation and the UAS mapped wetland. One hundred percent of the UAS mapped
wetland area is found within the delineated area, while the UAS mapped area covers
approximately 49.2% of the delineated area.
66

Figure 23. Case Study # 14 in Roosevelt, OK: (a) Esri base map with NWI mapped wetland; (b) UAS imagery with NWI mapped
wetland; (c) Wetland Delineation results, NWI mapped wetland and overlap between the two; (d) UAS mapped wetland results, NWI
mapped wetland and overlap between the two; (e) Overlap between Delineation and UAS mapped wetland results.
67

4.2.15. Case Study #15
Case Study #15 was flown on July 1, 2018 from 12:45 to 1:48 p.m. with a total flight
time of 1 hour and 3 minutes; this flight collected data for Case Studies #11 to #15. Weather
conditions were overcast skies with winds of ≤10 miles per hour. Delineation of the wetland took
place from 4:35 to 4:58 p.m. (i.e. 23 minutes). The wetland was surrounded by pasture land with
species such as honey mesquite, little bluestem, Indian grass, big bluestem, switchgrass, sideoats
grama, and bermuda grass. Wetland species included duckweed and blue green algae. Figure 24a
displays the Esri basemap, captured in July 2016, and the historic NWI mapped wetland (0.5
acres), captured in March 1983. There is a 33-year time lapse between these two data sets. Even
with the coarse imagery provided at the 1:1,000 scale, the NWI mapped area does not represent
the 2016 extent of the wetland. In Figure 24b the resolution is increased but the discrepancy
between the NWI data and visual estimation of the wetland location is still present. Figure 24c
shows the total delineated area (0.06 acres) and historic wetland overlap. When comparing
overall acreage of historic wetland versus the delineated area, the delineation method indicates
an 88% decrease in wetland size. Figure 24d shows the wetland area determined by supervised
classification via UAS imagery (0.01 acres) and historic wetland overlap. When comparing
overall acreage of the UAS versus the historic wetland, there is an 98% decrease in wetland size.
While the UAS collected data provides a more conservative estimate than a full delineation both
show a marked increase in wetland size when compared to the historic data captured 35 years
ago. Finally, Figure 24e shows the overlap between the wetland delineation and the UAS
mapped wetland. One hundred percent of the UAS mapped wetland area is found within the
delineated area, while the UAS mapped area covers approximately 16.7% of the delineated area.
68

Figure 24. Case Study # 15 in Roosevelt, OK: (a) Esri base map with NWI mapped wetland; (b) UAS imagery with NWI mapped
wetland; (c) Wetland Delineation results, NWI mapped wetland and overlap between the two; (d) UAS mapped wetland results, NWI
mapped wetland and overlap between the two; and (e) Overlap between Delineation and UAS mapped wetland results.
69

4.3 Cost-Benefit Analysis
The time required completing the processes for working with the wetland delineation and
the UAS were tallied separately and broken down into three categories, acquisition, processing,
and analysis. Table 10 shows the total time required for each category using the GPS receiver
and the UAS system in this study. The final tally shows a significant savings can be achieved
using the UAS. Mobilization time was not taken into account as travel would have been the same
for both data collection efforts. It is important to note that multiple wetlands were flown within a
single mission and only one wetland at a time could be assessed with traditional delineation
methods. Processing with Pix4D took place at night with limited input from the user after the
template was created and images uploaded, subsequently only the time used to upload the
imagery was considered for this comparison. The average salary of an entry level biologist with a
multiplier of three was used to estimate the costs. There was a considerable cost and time savings
when considering UAS against traditional methods. This cost analysis only considers the amount
of labor for a technician to perform the work and does not consider mileage during travel nor
equipment such as a soil auger, a Trimble GPS unit, vegetation field guides, rental rates from
consultants for UAS use, or software costs. The costs that were omitted typically vary between
consultants and/or project.
Table 10. Time required to acquire, process and analyze data collected for delineation and UAS.
Delineation UAS Cost of
Delineation
Cost of UAS
Acquisition 16 hr 30 min 3 hr 40 min $1,023 $248
Processing 5 hr 2 hr upload/
10 hr for Pix4D
output
$310 $124
Analysis 6 hr 7 hr $372 $434
Total 27 hr 30 min 20 hr 40 min $1,705 $806

70

Chapter 5: Discussion and Conclusions
This chapter discusses the broader significance and ramifications of the results of this
thesis and offers some conclusions about the use of UAS versus traditional delineation methods.
5.1 Findings
The methodology utilized in this study provided three significant findings: (1) wetland
determination is inherently flawed, utilizing data that was captured over 3 decades ago; (2) new
UAS methods offer ease of data acquisition in comparison to traditional methods; and (3)
substantial cost savings when compared to full delineation.
When one compares all of the maps from the case studies it can noted that the level of
resolution is a hindrance to the user when attempting to determine the location and extent of
wetlands or isolated ponds on the landscape using Esri basemaps. The lack of resolution, control
of flight conditions during image capture and the time gaps between Esri basemap capture and
project dates prevents consultants from having high confidence about the true condition of the
wetland. The NWI data was found to be 35-37 years old at the time of this thesis. Within each of
the 15 case studies that were provided not one NWI mapped wetland matched the delineated
wetland in size or placement. As seen with the confusion matrix results, the largest discrepancies
were found where dried soils were exposed that would have been inundated in wetter conditions.
Applying the buffer to the UAS mapped wetlands did not appear to drastically increase the
overlap or the percent accuracy. While some wetlands, such as Case Study #6, had similar shape
to current conditions the placement of the wetland was to the northwest while other wetlands
such as Case Study #11 are drastically different when considering the true extent of the wetland.
It was typically found that there was overlap with the delineated wetland and the NWI so if a true
71

delineation is required researchers would be able to know the approximate location of the
wetland if it still existed.
5.2 Advantages and Disadvantages of Using the UAS Methodology
The inexpensive UAS used to collect data for this study performed very well over water
bodies which have been noted to interfere with GPS signals (Matolak 2015). The UAS platform
is extremely user friendly and complete the data capture in a short period of time. It is also very
easy to mobilize in a safe location and observe the mission without leaving the area around the
vehicle. This limits the safety hazards that would be encountered during a traditional wetland
assessment. These hazards include but are not limited to dead or dying trees, entanglement in
thorny vegetation, potential biting insects that could carry disease (i.e. West Nile virus, Lyme
disease, rocky mountain spotted fever, etc.) and possible drowning.
The UAS estimation of wetland area was always a more conservative measurement than
what was found in a traditional wetland delineation. While the comparison of a delineation and
the UAS method is a comparison of two very different types of assessments, by assessing both
we were able to determine the accuracy of the UAS method, a comparison with the NWI would
have entailed comparing two datasets of unknown accuracy and error. The delineation method is
the most accurate assessment but is still subject to human error. The UAS method allowed for a
conservative method to increase placement accuracy on the landscape over the NWI method.
The average overlap between the delineation and UAS methods for wetlands located in Meeker,
Pawnee, and Roosevelt OK was 86.4, 87.8, and 70.6%, respectively. If Case # 15 is considered
an outlier, as there was a substantially large area that was dry pond bed, which brought the
overlap down to 16.7%, this would bring the Roosevelt site up to 76.6% overlap. While wetland
delineations are considered to be the highest quality determination method they are still subject
72

to human error and dependent on seasonal and/or yearly variability. The UAS on average was
typically found completely within the delineated area and the average overlap was 76.5% with
all sites and 80% excluding the outlier with the complete delineated area.
An unexpected development of flying missions within the growing season was the
presence of algae and pondweed floating near the edges and/or in the wetlands. The supervised
classification method was not able to determine the differences between the algae and the
surrounding grasslands. The supervised classification method was also not able to determine the
wetland extent underneath full canopy cover nor in areas that were desiccated from the original
boundary. The limitations of canopy cover/algae can be seen in Case Study #2 and the issues
with desiccated areas can be seen in Case Studies #12 and #15. To correct vegetation issues it
might be best to complete UAS missions during the leaf-off seasons. The following sections
explore corrections that could allow for flights to occur during the growing or leaf-off seasons.
5.2.1 Supplemental buffers
As seen in all case studies, the placement of the wetland was improved by the UAS
method even though this method was found to be more conservative than the full delineation.
The reason for these conservative estimates was attributed to dense canopy cover and floating
vegetation and to see if this could be corrected, 5, 10, 15, 20, 25, 30 and 35 ft buffers were
placed around the wetlands to determine if there was an optimal distance to use to encompass the
entire delineated area. By applying up to a 35 ft buffer on all sites 95% of the wetlands within
this study had full coverage. Table 11 shows the overlap percentage achieved with each buffer. It
can be seen that the 15 ft buffer encompasses most of the case studies with ≥ 90% overlap.
However, not all wetlands were completely encompassed even when using the 35ft buffer.
Because the UAS method was only assessing the open water areas of the wetland and the
73

vegetation was not incorporated in the supervised classification method applying a buffer
allowed for the surrounding plant life to be incorporated.
5.2.2 Seasonality
The season in which UAS missions occur could also increase the accuracy of the UAS
method. As seen in Table 3, only Case Study #5 was flown in a season that would be considered
leaf-off or non-growing (January 15, 2018). Figure 27 illustrates the difference between winter
and summer UAS flights and the overlap of the delineated area found in each season. The
Delineated Area of Case Study #5 was 0.87 acres, and the summer UAS mapped wetland
calculated that the total area of the wetland was 0.62 acres, with an overlap of 71.3%. In contrast
the winter flight allows the UAS to see through the tree canopy and determine the shoreline with
greater accuracy than was achieved in the summer months. The winter UAS mapped wetland
calculated the total area of the wetland to be 0.85 acres with an overlap of 97.7%. This improved
the overlap of the Delineated and UAS mapped wetlands by 26.4%. While this is only one
wetland, this could provide a key area of exploration for UAS wetland assessment in the future.
If winter months are to be considered for flights, it is suggested that they be performed in times
when the wetlands are not iced over as the classification of the wetlands could change due to
inclement conditions prior to or during the flights.
74

Table 11. Percent overlap achieved using at 5, 10, 15, 20, 25, 30, and 35 ft buffers generated around the 15 UAS mapped wetlands

Case Study Area
(ac)
5ft
(ac)
%
overlap
10ft
(ac)
%
overlap
15ft
(ac)
%
overlap
20ft
(ac)
%
overlap
25ft
(ac)
%
overlap
30ft
(ac)
%
overlap
35ft
(ac)
%
overlap
#1 1.47 1.39 94.56 1.43 97.28 1.46 99.32 1.47 99.93 1.47 99.96 1.47 99.96 1.47 99.96
#2 0.67 0.62 92.54 0.66 98.51 0.67 99.40 0.67 99.55 0.67 99.58 0.67 99.58 0.67 99.58
#3 0.35 0.35 100.00 0.35 100.00 0.35 100.00 0.35 100.00 0.35 100.00 0.35 100.00 0.35 100.00
#4 2.34 2.03 86.63 2.18 93.08 2.27 96.84 2.31 98.69 2.33 99.53 2.34 99.88 2.34 100.07
#5 0.87 0.71 82.05 0.78 89.73 0.81 93.68 0.84 95.98 0.85 97.23 0.85 97.92 0.86 98.38
#6 0.81 0.78 96.40 0.81 99.59 0.81 99.71 0.81 99.71 0.81 99.71 0.81 99.71 0.81 99.71
#7 0.87 0.75 86.44 0.77 88.27 0.78 89.91 0.80 91.59 0.81 93.29 0.83 95.03 0.84 96.64
#8 1.68 1.37 81.83 1.46 86.84 1.51 90.14 1.55 92.41 1.58 94.26 1.61 95.82 1.63 97.14
#9 0.86 0.80 93.30 0.83 96.17 0.84 98.00 0.85 99.29 0.86 99.99 0.86 100.32 0.86 100.40
#10 0.65 0.65 100.33 0.65 100.33 0.65 100.33 0.65 100.33 0.65 100.33 0.65 100.33 0.65 100.33
#11 0.67 0.64 95.43 0.66 99.01 0.67 99.34 0.67 99.34 0.67 99.34 0.67 99.34 0.67 99.34
#12 0.23 0.13 58.68 0.15 67.34 0.17 73.29 0.18 78.20 0.19 82.52 0.20 86.30 0.21 89.47
#13 0.27 0.24 90.30 0.27 98.90 0.27 100.27 0.27 100.27 0.27 100.27 0.27 100.27 0.27 100.27
#14 0.59 0.37 62.77 0.43 72.09 0.46 78.66 0.49 83.46 0.51 86.52 0.53 89.27 0.54 92.05
#15 0.06 0.01 24.72 0.03 45.54 0.04 64.44 0.05 78.68 0.05 87.33 0.06 92.09 0.06 93.92
75

Some USACE districts require that delineation take place during the growing season, so
if a comparison was to be made to meet both this USACE requirement and the preference for
leaf-off time periods early spring might be the better option. Lovvorn and Kirkpatrick (1982)
found that aerial photographs have potential for the accurate identification of all dominant
species and that early September is an optimal time for species identification due to the variation
in fall colors among species. However, Tiner (1990) determined that CIR aerial photography in
early spring is best for detecting deciduous forested wetlands in temperate regions. Due to this
conflicting advice in the literature, research will be required for different regions to determine
the optimal seasons for image capture. Buffer distances do not take into account the topography
of the land surface and this leads to misperceptions in and near depressions. This causes non-
uniform growth of wetlands in certain directions that cannot be handled by buffering alone.
Figure 25 shows the success of this method with Case Study # 11 which had little canopy cover
and therefore needed a limited buffer. Figure 26 shows Case Study # 7 which has an oddly
shaped wetland and the full delineated area was not covered.
76

Figure 25. A map of Case Study #11 showing how the application of a buffer around the UAS mapped wetlands allowed for
the delineated wetland to be fully encompassed within a 15 ft buffer.
77

Figure 26. A map of Case Study #7 show how the application of a buffer around the UAS mapped wetlands which does not
consider the topography with a 35 ft buffer.
78

Figure 27. Differences in overlap between winter and summer UAS flight times and the delineated wetland for Case Study #5.
79

5.2.3 Other potential methods
Other methods that have potential for assessment of wetlands include the digital surface
models that are provided in programs such as Pix4D. UASs are distinctive in this regard because
satellite images do not provide these types of returns, and LiDAR technology, while accurate, is
typically not cost effective for many projects. To collect DSMs, missions are typically flown in a
3D mapping grid in order to get a better understanding of the landscape. This would be best
performed on areas that do not have high percentages of canopy cover. A digital surface model
has elevations that are represented with the first reflected surface detected by the sensor. These
first returns may be reflected by bare ground or by surface features such as trees and structures.
Figures 28 and 29 show an example of how digital surface models could be used to
predict and visualize the directional growth in wetlands. Within these maps one can see the
current extent of the wetland and the potential for directional growth to the north and east.
Information like this provides a more accurate assessment of the site than additional buffers or
different seasons.
Other potential areas of exploration would be the investigation of thermal imagery. Water
tends to have a cooling effect on surrounding substrates, one could hypothesize that if flying with
a thermal camera, areas that are completely saturated or in close proximity to those areas would
be cooler than areas without water. However, these missions would have varying results due to
the time of year and vegetation cover, similar to other methods.
80

Figure 28. Slope estimate from Digital Surface Model provided in Pix4D with current
wetland extent.

Figure 29. Wetland growth estimate from Digital Surface Model provided in Pix4D with
current wetland extent.

81

5.3 Conclusions
Wetlands are dynamic systems with seasonal/yearly changes as seen with the rainfall
differences between image capture dates. Even wetland delineations are subject to human error
and the problems caused by limitations in accessibility. The increase in placement and size
accuracy of the wetlands using UAS has drastically improved upon the current datasets being
utilized for wetland assessment. The UAS mapped wetlands, on average, were located
completely within the delineated area and the average overlap was 76.5% with all sites and 80%
excluding the outlier with the complete delineated area. UAS was typically a more conservative
estimate of area than a full wetland delineation. By adding up to a 35 ft buffer the UAS mapped
wetlands typically encompass ≥90% of the delineated wetland. UAS is limited by canopy cover,
floating vegetation, and most severely the presence of dry wetland beds and additional methods
such as including buffers, seasonality of missions, assessments of topography from digital
surface models, and thermal imagery have been suggested as areas of future research to build
upon this work. The cost-savings of this method when compared to traditional methods reduced
costs by $1,000 for labor alone; in large projects this could result in significant savings to
consultants and project proponents. This method could be used as a pre-planning tool for wetland
avoidance during development; however, as technological advancement continues traditional
methods may well be used sparingly in favor of the rapid and cost-effective methods provided by
UAS.

82

REFERENCES
Abd-Elrahman, A., L. Pearlstine, F. Percival. 2005. “Development of Pattern Recognition
Algorithm for Automatic Bird Detection from Unmanned Aerial Vehicle Imagery.”
Surveying and Land Information Science 65: 37-45
Adam, E., Mutanga, O., and Rugege, D. 2010. “Multispectral and hyperspectral remote sensing
for identification and mapping of wetland vegetation: a review.” Wetlands Ecology
Management 18(3): 281-296.
Ahmad, A., K.N. Tahar, W.S. Udin, K.A. Hashim, N.H. Darwin, M. Hafis, M. Room, N.F.A.
Hamid, N.A.M. Azhar, and S.M. Azmi. 2013. “Digital aerial imagery of unmanned aerial
vehicle for various applications.” In Control System, Computing and Engineering
(ICCSCE), 2013 IEEE International Conference on, 535-540.
Berteska, T., and B. Ruzgiene. 2013. “Photogrammetric mapping based on UAS imagery.”
Geodesy and Cartography 39 (4): 158.
Breckenridge, R.P., and Dakins, M.E. 2011. “Evaluation of bare ground on rangelands using
unmanned aerial vehicles: a case study.” GIScience and Remote Sensing 48(1): 74-85.
Cai, Y.; Sun, G.; Liu, B. 2005. “Mapping of water body in poyang lake from partial spectral
unmixing of modis data.” In Proceedings of the 2005 IEEE International Geoscience
and Remote Sensing Symposium (IGARSS ’05), Seoul, Korea 4539–4540.
Campbell, J. B., and R.H. Wynne. 2011. “Introduction to remote sensing.” Fifth ed. New York,
NY: The Guilford Press.
Chabot, Dominique, and David M. Bird. 2013. "Small unmanned aircraft: precise and
convenient new tools for surveying wetlands." Journal of Unmanned Vehicle Systems,
pg. 15.
Dale, P.E.R.; Hulsman, K.; Chandica, A.L. 1986. “Classification of reflectance on colour
infrared aerial photographs and sub-tropical salt-marsh vegetation
types.” International Journal of Remote Sensing 7: 1783–1788.
Dunfford, R., Michel, K., Gagnage, M., Piegay, H., and Tremelo, M.-L. 2009. “Potential and
constraints of unmanned aerial vehicle technology for the characterization of
Mediterranean riparian forest.” International Journal of Remote Sensing 30(19): 4915-
4935.
Dustin, M.C. 2015. Monitoring Parks with inexpensive UASs: Cost Benefits analysis for
monitoring and maintain park facilities. Thesis, University of Southern California.
83

Draeyer, B., Strecha, C. (2014) Pix4D White paper: How accurate are UAS surveying
methods? Internet: https://support.pix4d.com/hc/en-us/articles/202557229-Scientific-
White-Paper-How-accurate-are-UAS-surveying-methods-gsc.tab=0 (16.02.2017)
Everitt, J. H., C. Yang, F.W. Judd, and K.R. Summy. 2010. “Use of archive aerial photography for
monitoring black mangrove populations.” Journal of Coastal Research 26(4): 649- 53.
Everitt, J.H.; Yang, C.; Summy, K.R.; Judd, F.W.; Davis, M.R. 2014. “Evaluation of color-
infrared photography and digital imagery to map black mangrove on the Texas Gulf
Coast.” Journal of Coastal Research 23: 230–235.
Everitt, J.H.; Yang, C.; Judd, F.W.; Summy, K.R. 2015. Use of archive aerial photography for
monitoring black mangrove populations. Journal of Coastal Research 26: 649–653.
Falkner, E., and D. Morgan. 2002. “Chapter 17 project planning and cost estimating.” Aerial
mapping methods and applications. Second ed., 161-172 CRC Press LLC.
Finlayson, C., & Mitchell, D. 1999. “Australian wetlands: The monitoring challenge. Wetlands
Ecology and Management.” 7(1-2): 105-112.
Gademer, A., F. Mainfroy, L. Beaudoin, L. Avanthey, V. Germain, C. Chéron, S. Monat, and J.P.
Rudant. 2009. “Solutions for near real time cartography from a mini-quadrotor UAS.” In
SPIE Europe Remote Sensing.” International Society for Optics and Photonics. pp.
74781G-74781G.
Getzin, S., Wiegand, K., and Schoning, I. 2012. “Assessing biodiversity in forests using very
high-resolution images and unmanned aerial vehicles methods.” Ecological Evolution
3(2): 397-404.
Green, E.P.; Ellis, A.C. 1998. “The assessment of mangrove areas using high resolution
multispectral airborne imagery.” Journal of Coastal Research 14: 433–443.
Guo, M.; Li, J.; Sheng, C.; Xu, J.; Wu, L. 2017 “A Review of Wetland Remote
Sensing.” Sensors 17: 777.
Hardin, P.J., and Jackson, M.W. 2005. An unmanned aerial vehicle for rangeland photography.
Rangeland Ecology and Management 58(4): 439-442.
Jones IV, G. P., L.G. Pearlstine, and H.F. Percival. 2006. “An assessment of small unmanned
aerial vehicles for wildlife research.” Wildlife Society Bulletin 34 (3): 750-8.
Johannessen, C.L. 1964. “Marshes prograding in Oregon: Aerial photographs.” Science 146:
1575–1578.
84

Laliberte, A.S., A. Rango, and J.E. Herrick. 2007. “Unmanned aerial vehicles for rangeland
mapping and monitoring: a comparison of two systems.” Proceedings of the ASPRS
Annual Conference, May 7-11, Tampa, FL.
Lee, K.H. and Lunetta, R.S. 1995. “Wetland detection methods. In Wetland and Environmental
Application of GIS” Lewi Publishers, New York, NY. Pp. 249-284.
Lovvorn, J.R. and Kirkpatrick, C.M. 1982. “Analysis of freshwater wetland vegetation with
large-scale color infrared aerial photography.” Journal of Wildlife Management 46: 61–
70.
Matolak, J.D. 2015. “Unmanned aerial vehicles: Communications challenges and future aerial
networking”, 2015 International Conference on Computing, Networking, and
Communication, vol. 1, pp. 567-572, Feb. 2015
OKGOV. 2018. Wetland Mapping. Accessed January 31, 2018. Available online at:
https://www.ok.gov/wetlands/Assessment/Wetland_Mapping/index.html
Ordoyne, C.; Friedl, M.A. 2008. “Using MODIS data to characterize seasonal inundation
patterns in the Florida Everglades.” Remote Sensing of the Environment 112: 4107–
4119.
Paine, D. P., and J. D. Kiser. 2012. “Aerial photography and image interpretation third edition.”
Third ed. Hoboken, New Jersey: John Wiley & Sons, Inc.
Rango, A., Laliberte, A., Herrick, J.E., Winters, C., Havstad, K., Steele, C., and Browning, D.
2009. “Unmanned aerial vehicle-based remote sensing for rangeland assessment,
monitoring, and management.” Journal of Applied Remote Sensing 3: 335-342
Rango, A., A.S. Laliberte, C. Steele, J.E. Herrick, B. Bestelmeyer, T. Schmugge, A. Roanhorse,
and V. Jenkins, 2006. “Using unmanned aerial vehicles for rangelands: current
applications and future potentials.” Environmental Practice 8: 159-168
Rogers, D.J.; Hay, S.I.; Packer, M.J.; Wint, G.R.W. 1997. “Mapping land-cover over large areas
using multispectral data derived from the NOAA-AVHRR: A case study of
Nigeria.” International Journal of Remote Sensing 18: 3297–3303.
Scarpace, F.L.; Quirk, B.K.; Kiefer, R.W.; Wynn, S.L. 1982. “Wetland mapping from digitized
aerial photography.” Photogrammetric Engineering and Remote Sensing 47: 829–838.
Tiner, R.W. 1990. “Use of high-altitude aerial photography for inventorying forested
wetlands in the United States.” Forest Ecology and Management 33: 593–604.
85

USACE. 1987. Corps of Engineers Wetlands Delineation Manual. Environmental Laboratory
U.S. Army Corps of Engineers, Waterways Experiment Station, Wetlands Research
Program Technical Report Y-87-1. Vicksburg, MS.
USACE. 2010. “Regional Supplement to the Corps of Engineers Wetland Delineation Manual:
Great Plains Region (Version 2.0)”
USEPA. 2002. Methods for Evaluating Wetland Condition: Wetlands Classification. Office of
Water, U.S. Environmental Protection Agency, Washington, DC. EPA-822-R-02-017.
USEPA. 2017. Section 404 of the Clean Water Act. Accessed January 30 2018. Available online at:
https://www.epa.gov/cwa-404/section-404-permit-program
USFWS 2018a. Data Limitations, Exclusions and Precautions. Accessed: January 30 2018.
Available online at: https://www.fws.gov/wetlands/data/Limitations.html
USFWS. 2018b. National Wetlands Inventory website. U.S. Department of the Interior, Fish and
Wildlife Service, Washington, D.C. http://www.fws.gov/wetlands/
Trimble Navigation Limited. Pathfinderoffice. 2015 [cited April 10 2015]. Available from
http://www.trimble.com/mappingGIS/PathfinderOffice.aspx.
Watts A.C, V.C. Ambrosia, and E.A. Hinkley 2012. “Unmanned aircraft systems in remote
sensing and scientific research: classification and considerations of use.” Remote Sensing
4: 1671–92.
Watkins, S., J. Milbank, B.J. Loxton, and W.H. Melbourne. 2006. “Atmospheric winds and their
implications for microair vehicles.” AIAA Journal 44(11): 2591-600.
Woods, A.J., Omernik, J.M., Butler, D.R., Ford, J.G., Henley, J.E., Hoagland, B.W., Arndt, D.S.,
and Moran, B.C., 2005, Ecoregions of Oklahoma (color poster with map, descriptive text,
summary tables, and photographs): Reston, Virginia, U.S. Geological Survey (map scale
1:1,250,000).
Xiang, H. and L. Tain. 2011. “Method for automatic georeferencing aerial remote sensing (RS)
images from an unmanned aerial vehicle (UAS) platform.” Biosystems Engineering
108(2): 104-113.
Zhang, L., Wang, M.-H., Hu, J., Ho,Y.-S. 2010. “A review of published wetland research, 1991-
2008: Ecological engineering and ecosystem restoration.” Ecological Engineering 36: 973-
980.

Abstract (if available)

Abstract When project proponents wish to assess a development site for jurisdictional wetland impacts, they are traditionally left with two options: a wetland determination or a delineation. A wetland determination is customarily a desktop assessment of the site including, but not limited to, the following datasets: National Wetland Inventory (NWI), National Hydrography Dataset (NHD), National Resources Conservation Services Soil Survey (NRCS), topographic maps and satellite imagery. A wetland delineation assesses the presence of hydrophytic vegetation, hydric soils and hydrology during field evaluation. The NWI is typically used to determine where existing wetlands are in order to determine if they qualify as jurisdictional wetlands. This allows project proponents to either take the appropriate avoidance measures to reduce impacts to the wetland or determine if a full wetland delineation is required to apply for a Section 404 permit. In some cases, NWI maps have not been updated for up to 30 years, and these mapped wetlands are limited by conditions that were present at the time the aerial imagery was taken. ❧ This thesis shows that by incorporating unmanned aerial systems (UAS) into a wetlands determination, wetland specialists and project planners can capture current conditions of the development site (i.e. topography, disturbance, land cover, etc.) within efficient time frames and assess the potential extent of a wetland(s). This allows project proponents to avoid the cost and time restrictions that come from a full wetland delineation. The UAS imagery was compared to historically mapped wetlands still present

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Utilizing advanced spatial collection and monitoring technologies: surveying topographical datasets with unmanned aerial systems

PDF

Wetland mapping and restoration decision making using remote sensing and spatial analysis: a case study at the Kawainui Marsh

PDF

Data overload in unmanned aircraft systems: improving bandwidth utilization through wavelet compression

PDF

An application of aerial drones in high definition mapping for autonomous vehicles

PDF

Monitoring parks with inexpensive UAVs: cost benefits analysis for monitoring and maintaining parks facilities

PDF

Unmanned aerial systems for surveying and mapping: cost comparison of UAS versus traditional methods of data acquisition

PDF

Utilizing online data sources to improve existing military aircraft systems

PDF

Urban areas and avian diversity: using citizen collected data to explore green spaces

PDF

Surface rupture detection with support vector machine classification: case study from Ridgecrest, CA

PDF

Historical ecology of the Split Oak Forest in east central Florida

PDF

A comparison of GLM, GAM, and GWR modeling of fish distribution and abundance in Lake Ontario

PDF

Using aerial imagery to assess tropical forest cover surrounding restoration sites in Costa Rica

PDF

Evaluating machine learning tools for humanitarian road network mapping

PDF

A spatial investigation of New York City's historical shoreline

PDF

Filling in the gaps: 3D mapping Arizona's Basin and Range aquifer in the Prescott Active Management Area

PDF

Modeling burn probability: a Maxent approach to estimating California's wildfire potential

PDF

Exploring remote sensing and geographic information systems technologies to understand vegetation changes in response to land management practices at Finke Gorge National Park, Australia Between ...

PDF

Assessing woody plant encroachment in Marin County, California, 1952-2018

PDF

Walkability study for school accessibility: case study of the San Juan, Puerto Rico elementary schools

PDF

Evaluating the utility of a geographic information systems-based mobility model in search and rescue operations

Asset Metadata

Creator Burchette, Monika Lynne (author)

Core Title Improving wetland determination utilizing unmanned aerial systems

School College of Letters, Arts and Sciences

Degree Master of Science

Degree Program Geographic Information Science and Technology

Publication Date 10/15/2018

Defense Date 08/10/2018

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag DJI Phantom,OAI-PMH Harvest,supervised classification,UAS,UAV,unmanned aerial systems,wetland determination,wetlands

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Wilson, John (committee chair), Longcore, Travis (committee member), Marx, Andrew (committee member)

Creator Email mburchet@usc.edu,mburchette@olssonassociates.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c89-77414

Unique identifier UC11672290

Identifier etd-BurchetteM-6829.pdf (filename),usctheses-c89-77414 (legacy record id)

Legacy Identifier etd-BurchetteM-6829.pdf

Dmrecord 77414

Document Type Thesis

Format application/pdf (imt)

Rights Burchette, Monika Lynne

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA