University of Southern California Dissertations and Theses

From ruins to pixels: using remote sensing and GIS to analyze, document, and visualize archaeological sites and ancient roadways in Chaco Canyon, New Mexico

(USC Thesis Other)

From ruins to pixels: using remote sensing and GIS to analyze, document, and visualize archaeological sites and ancient roadways in Chaco Canyon, New Mexico

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content FROM RUINS TO PIXELS: USING REMOTE SENSING AND GIS TO ANALYZE,
DOCUMENT, AND VISUALIZE ARCHAEOLOGICAL SITES AND ANCIENT
ROADWAYS IN CHACO CANYON, NEW MEXICO
by
Paul Penna
A Thesis Presented to the
FACULTY OF THE USC DORNSIFE COLLEGE OF LETTERS, ARTS AND SCIENCES
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
MASTER OF SCIENCE
(GEOGRAPHIC INFORMATION SCIENCE AND TECHNOLOGY)
December 2024
Copyright 2024 Paul Penna

ii
To my parents, my Grammy, Florencia, and my puppy Bear

iii
Acknowledgements
I am deeply grateful to my mentor, Dr. Yi Qi, for providing me with invaluable guidance and
direction throughout my research project. His advice and expertise have been instrumental in
shaping my research and helping me to overcome various challenges along the way.
Additionally, I want to thank all of the other faculty members from SSI who have generously
offered their time and assistance whenever I needed it. I also express my sincere appreciation to
Land Info Worldwide Mapping and OpenTopography for providing me with access to their
extensive data resources. Their data has been an integral part of my research project and has
allowed me to gain valuable insights into my topic. Finally, I would like to thank the National
Parks Service, specifically the supervisory archaeologist at Chaco Canyon, Lori Stephens, for her
expertise, assistance and access to necessary park resources.

iv
Table of Contents
Dedication...................................................................................................................................... iv
Acknowledgements........................................................................................................................ iii
List of Figures................................................................................................................................ vi
Abbreviations............................................................................................................................... viii
Abstract.......................................................................................................................................... ix
Chapter 1 Introduction .................................................................................................................... 1
1.1 Background ......................................................................................................................... 3
1.2 Study Area........................................................................................................................... 3
1.2.1 Chaco Canyon National Historical Park .................................................................... 3
1.2.2 Site Bc 53................................................................................................................... 5
1.3 The Anasazi and Chaco Canyon......................................................................................... 6
1.3.1 Historical Description of Area ................................................................................... 7
1.3.2 Site Bc 53 Background .............................................................................................. 8
1.3.3 Anasazi Culture and Society.................................................................................... 10
1.3.4 The Ancient Chacoan Roadway Network................................................................ 12
1.3.5 Descendants of the Anasazi and Chaco Canyon’s Role in Modern Culture............ 14
1.4 Project Overview............................................................................................................... 17
1.5 Document Overview ......................................................................................................... 19
Chapter 2 Related Work................................................................................................................ 20
2.1 Supervised Image Classification ....................................................................................... 20
2.1.1 Previous Archaeological Research Employing Image Classification...................... 21
2.1.2 Foundational Literature on Supervised Image Classification in Archaeology ........ 22
2.2 Site Bc 53 and Least Cost Path Analysis.......................................................................... 22
2.2.1 Previous Research Regarding Site Bc 53................................................................. 23
2.2.2 Critical Literature for Site Bc 53 and LCP Analysis ............................................... 25
2.3 3D Visualization of Chaco Canyon .................................................................................. 27
2.3.1 Visualizing Chaco Canyon and Similar Archaeological Sites in 3D....................... 27
2.3.2 Fundamental Archaeological Research Employing 3D Modeling and LiDAR....... 28
Chapter 3 Research Design and Methodology.............................................................................. 31
3.1 Overview........................................................................................................................... 31
3.2 Data................................................................................................................................... 31
3.3 Exploratory Analysis Methodology.................................................................................. 35
3.4 Site Bc 53 Analysis Methodology .................................................................................... 37
3.5 3D Visualization Methodology......................................................................................... 40
Chapter 4 Results.......................................................................................................................... 44
4.1 Supervised Image Classification ....................................................................................... 44
4.1.1 Random Trees.......................................................................................................... 44

v
4.1.2 Support Vector Machine .......................................................................................... 47
4.2 Significance of Site Bc 53................................................................................................. 50
4.2.1 LCP Analysis Between Site Bc 53 and the Nearby Casa Rinconada Structures..... 51
4.2.2 LCP Analysis Between Site Bc 53 and the Nearby FP Chacoan Staircase ............. 53
4.3 Assessment of the Physical Condition of Site Bc 53........................................................ 55
4.3.1 Assessing Site Bc 53 Using Supervised Image Classification Results.................... 55
4.3.2 Assessing Site Bc 53 Using LiDAR-Derived DEM................................................ 56
4.4 3D Visualization of the LiDAR-Derived Digital Elevation Model.................................. 59
4.4.1 Chaco Canyon Visualized in a 3D Scene in ArcGIS Pro ........................................ 60
4.4.2 Combining the 3D Visualization with High-Resolution Satellite Imagery ............. 61
4.4.3 Ethical Implications of the Final 3D Visualization.................................................. 67
Chapter 5 Conclusions and Discussion ......................................................................................... 68
5.1 Deciphering the Results.................................................................................................... 69
5.1.1 Supervised Image Classification Results................................................................. 69
5.1.2 Site Bc 53 Analysis Breakdown .............................................................................. 70
5.1.3 3D Visualization Product and Results..................................................................... 74
5.2 Evaluating the Overall Research and Future Directions................................................... 75
5.2.1 Research Limitations................................................................................................ 75
5.2.2 Future Research........................................................................................................ 76
References..................................................................................................................................... 78

vi
List of Figures
Figure 1. Chaco Canyon National Historical Park.......................................................................... 5
Figure 2. Study Area 2, Site Bc 53 ................................................................................................. 6
Figure 3. Google Earth view of Chaco Canyon............................................................................ 32
Figure 4. Raw Multispectral Satellite Imagery ............................................................................. 33
Figure 5. Raw LiDAR-derived DEM............................................................................................ 34
Figure 6. Exploratory Analysis Workflow.................................................................................... 37
Figure 7. Site Bc 53 Analysis Workflow ...................................................................................... 40
Figure 8. 3D Visualization Workflow .......................................................................................... 42
Figure 9. 3D visualization uploaded to ArcGIS Online................................................................ 43
Figure 10. Satellite imagery at 30 cm resolution from Pleiades Satellite ..................................... 45
Figure 11. Classification of Casa Rinconada employing the RT Algorithm ................................ 46
Figure 12. Random Trees Algorithm Confusion Matrix .............................................................. 47
Figure 13. Satellite imagery at 30 cm resolution from Pleiades Satellite ..................................... 48
Figure 14. Classification of Pueblo Bonito employing SVM Algorithm ..................................... 49
Figure 15. Support Vector Machine Algorithm Confusion Matrix .............................................. 50
Figure 16. Points for the LCP Analysis and LiDAR-derived DEM input raster.......................... 51
Figure 17. Least cost path between Bc 53 and Casa Rinconada Community............................... 52
Figure 18. Least cost path between Bc 53 and the FP Chacoan Staircase.................................... 54
Figure 19. Supervised image classification (SVM) of Bc 53 ....................................................... 56
Figure 20. LiDAR-derived DEM of Bc 53 and Casa Rinconada Community ............................. 57
Figure 21. LiDAR-derived DEM of the Casa Rinconada Community......................................... 58
Figure 22. LiDAR-derived DEM of Bc 53................................................................................... 59

vii
Figure 24. Generated 3D visualization of Chaco Canyon facing Pueblo Bonito ......................... 61
Figure 25. Generated 3D visualization of Bc 53 and Casa Rinconada Community..................... 62
Figure 26. LCP analysis results displayed in the 3D visualization ............................................... 63
Figure 27. Final 3D visualization of Chaco Canyon zoomed-out ................................................ 65
Figure 28. Short distance ancient roadway network..................................................................... 66
Figure 29. Long distance ancient roadway network..................................................................... 67

viii
Abbreviations
ALS Airborne laser scanning
GIS Geographic information system
LCP Least cost path
LiDAR Light detection and ranging
RT Random trees algorithm
SSI Spatial Sciences Institute
SVM Support vector machine algorithm

ix
Abstract
This thesis project focuses on using advanced remote sensing technology to create a
comprehensive 3D model of Chaco Canyon, a significant US National Historical Park and
UNESCO World Heritage Site located in New Mexico. There are two primary motivations for
this research. The first motivation is that a contiguous 3D Visualization of Chaco Canyon that is
available for public viewing does not yet exist. The other primary motivation is to explore an
under-documented archaeological site within Chaco Canyon that is vulnerable to becoming fully
deteriorated, according to the results of this research. The project is built upon three interrelated
sub-research objectives. First, a supervised image classification is conducted on satellite imagery
of Chaco Canyon, which is then utilized to identify new archaeological areas of interest based on
the spectral signatures of the known ruins and roadways. Second, LiDAR data is used to
investigate a concealed site, and then an LCP analysis is performed to model potential travel
routes to nearby ruins from the concealed site. To conclude, a 3D visualization of Chaco Canyon
is generated using LiDAR point cloud data and high-resolution satellite imagery. By employing
cutting-edge remote sensing techniques and GIS methodologies, the product of this project is a
contiguous 3D visualization of Chaco Canyon that displays the locations of the major ruins and
illustrates the extent of the ancient roadway network. Additionally, this study seeks to support
ongoing cultural heritage preservation efforts at Chaco Canyon. The findings will benefit the
National Parks Service, associated tribes, conservation groups, and the broader academic and
public communities by providing a complete 3D visualization that can be used for educational
purposes, preservation efforts, informing public policy, and as a foundation for future
archaeological research at Chaco Canyon.

1
Chapter 1 Introduction
This research project employs a three-pronged approach to achieve its objectives. The first phase
of this research is to conduct a supervised image classification on satellite imagery obtained from
the Pléiades satellite constellation (acquired from Land Info Worldwide Mapping), which offers
a spatial resolution of 30 cm. The findings from the supervised image classification are utilized
to identify the location of ancient roadways and hidden structures that the desert landscape may
conceal by first defining the spectral signatures of the documented ruins and roadways and then
identifying any spectral similarities within the imagery.
The next phase of this research project is the investigation into a concealed structure on
the fringes of the canyon. For this step, the history of the hidden structure, called site Bc 53 or
Roberts’ Site, is examined. This step aims to build upon the supervised image classification
completed in phase one and to specifically determine whether the spectral signature of site Bc 53
matches the spectral signature of the nearby documented ruins, which could indicate a shared
building material and a possible community connection. The next step in this second phase is to
use the supervised image classification results again to assess the overall physical condition of
site Bc 53 in comparison to the condition of the other nearby ruins. This second research phase,
focusing on site Bc 53, concludes with a LCP analysis investigating a potential connection
between site Bc 53 and both the nearby FP Chacoan Staircase and Casa Rinconada community
of structures. The least-cost path analysis results provide insights into potential travel routes that
ancient inhabitants of Chaco Canyon may have utilized. This analysis also seeks to clarify if site
Bc 53 and the Casa Rinconada community are two separate areas or if they should all be
classified as the Casa Rinconada community. The information gained from this LCP analysis

2
enriches our understanding of ancient travel patterns, social dynamics, and landscape utilization
within the Chaco Canyon region and the American Southwest.
The third and final phase of this research project is to generate a 3D visualization using
LiDAR point cloud data and high-resolution satellite imagery. The satellite imagery is wrapped
onto the “blank” 3D model to create a realistic 3D reconstruction of the Chaco Canyon sites and
ancient roadways that radiate from them. The roadways and structures already documented in the
previous Chaco Canyon literature are integrated into the final 3D visualization (Friedman,
Sofaer, and Weiner 2021). Additionally, the potential roadways identified during the image
classification and the LCP analysis for this project are also illustrated in the 3D visualization.
The 3D visualization is made available for public viewing on ArcGIS Online.
By using a multi-layered research approach for this project, including 3D reconstruction,
image classification, LCP analysis, and cartographic visualization, this research seeks to uncover
new insights regarding the advanced trade, communication, and societal dynamics of the
Ancestral Pueblo. The findings of this research contribute to the ongoing efforts of cultural
heritage preservation at Chaco Canyon by emphasizing the importance of protecting these
archaeological sites from modern threats, such as encroaching drilling and gas production in the
area. Understanding the significance of these ruins and the intricacies of the roadway network is
crucial for preserving the cultural heritage at Chaco Canyon. This research project builds upon
the previous work conducted on the Ancestral Pueblo at Chaco Canyon by cartographically
documenting the ancient structures and roadway network in 3D and making it available for
public viewing online.

3
1.1 Background
Chaco Canyon is a US National Historical Park and UNESCO World Heritage Site with
an abundance of ancient structures and complex roadways built by the Ancestral Puebloans, also
known as the Anasazi, of northwestern New Mexico between 850 and 1200 CE. The overarching
topic of investigation is a three-tiered methodology that first conducts a supervised image
classification, next examines a specific ruin using an LCP analysis and LiDAR data, and finally
generates a 3D visualization of Chaco Canyon in its entirety. This thesis project utilizes
advanced GIS and remote sensing technology to complete the necessary analyses and generate
the expected final products. This research study concluded by generating a realistic 3D
visualization that is publicly available and can be used for a wide range of useful applications.
The primary data utilized to complete this research are the LiDAR-derived point cloud of the
Chaco Canyon study area and satellite imagery of Chaco Canyon at both the 1-meter and 30 cm
spatial resolutions.
1.2 Study Area
This section provides the current physical descriptions of the two study areas for this
research. The first study area is Chaco Canyon National Historical Park, which is the
encompassing study area for this research project. The second study area is a specific area within
Chaco Canyon National Historical Park, called site Bc 53.
1.2.1 Chaco Canyon National Historical Park
Chaco Canyon National Historical Park, located in northwestern New Mexico, is home to
an abundance of significant archaeological sites, and it is known for its continuous connection to
the Ancestral Chacoan culture. Chaco Canyon is situated within the Chaco Canyon National
Historical Park, encompassing an area of approximately 138 km2.

4
As displayed in Chaco Canyon lies within a geological setting characterized by deep
sandstone canyons, mesas, and arid desert landscapes. The Chaco Wash, as displayed in Figure
1,which is a seasonal stream that flows intermittently through the region, has carved this canyon.
The elevation of Chaco Canyon ranges from around 1,890 to 1,980 meters above sea level. The
canyon is surrounded by a dramatic landscape of red rock formations, including mesas and
buttes, which add to its scenic beauty and archaeological significance. Despite its remote
location, Chaco Canyon is accessible by road, although the last several miles leading to the park
are unpaved. The climate in Chaco Canyon is semi-arid, characterized by hot summers and cold
winters. Annual precipitation is relatively low, averaging around about 20 to 25 centimeters per
year, with most rainfall occurring during the summer monsoon season. The flora and fauna of
Chaco Canyon are adapted to the arid environment and include species such as piñon pine,
juniper, and sagebrush. Wildlife in the area includes mule deer, coyotes, rabbits, and numerous
bird species. Chaco Canyon is protected as part of the Chaco Culture National Historical Park,
which was established in 1907 to preserve its archaeological and cultural resources. The National
Park Service manages the park and encompasses both the canyon itself and surrounding areas of
cultural and natural importance. Chaco Canyon is significant because it is designated as both a
US National Historical Park and a UNESCO World Heritage Site; Chaco Canyon is recognized
internationally for its cultural and historical importance. Its inclusion on these prestigious lists
underscores its significance as a site of global human heritage.

5
1.2.2 Site Bc 53
As part of this thesis project, a specific area within Chaco Canyon called site Bc 53 or
Roberts' Site is examined in both spatial and societal contexts. As shown in Figure 2, site Bc 53
is located on the south side of the canyon, at the base of a cliff, and is situated across the Chaco
Wash from Pueblo Bonito. It is positioned 0.33 km east of Casa Rinconada and the cluster of
houses that make up the Casa Rinconada community. Bc 53 is comprised of approximately 21
rooms and four Kivas and was excavated by the University of New Mexico and School of
American Research field schools in 1940-1941 under the direction of Frank H. H. Roberts and
Paul Reiter. The structure started small and was built larger over a long period of time. Based on
the ceramic frequencies found at the site, which primarily consisted of Exuberant Corrugated,
Chaco, and Escavada Black-on-white ceramics, it is suggested that the occupation primarily
Figure 1. Study Area 1, Chaco Canyon National Historical Park

6
occurred during the last half of the 11th century. Bc 53 is located about 137 meters northeast of
Bc 51 and approximately 77 meters southwest of the FP Chacoan Staircase.
Figure 2. Study Area 2, Site Bc 53
1.3 The Anasazi and Chaco Canyon
This section provides the historical context for the research. It includes a description of
the landscape and how the Ancestral Puebloans worked with the land and created their homes
within the canyon. The section also gives background information about the Anasazi society and
their relationship to Chaco Canyon. It concludes by discussing the modern descendants of the
Anasazi and Chaco Canyon's role in modern-day culture.

7
1.3.1 Historical Description of Area
Chaco Canyon was the center of Ancestral Puebloan culture between the 9th and 12th
centuries CE, serving as a hub for trade, ceremony, and cultural exchange. The ancient Chaco
Canyon landscape was very similar to the climate in the area today, with long and brutal winters
that slowly transformed into scorching summertime temperatures. Also similar to today’s
climate, the Canyon from the 9th to the 12th centuries was categorized by short agricultural
seasons and marginal rainfall (Oswald 2018). While the consistent air desert climate over the
past one thousand years has protected many of the ancient structures created by the Anasazi, due
to climate change as a result of human behaviors, the area now is experiencing more extreme
weather than ever before, leaving the surviving structures and cultural material of Anasazi
vulnerable to rapid deterioration. The river or arroyo that flows through the region is called the
Chaco Wash and is the same water source that the Anasazi people relied on between the 9th and
12th centuries. While the modern Chaco Wash is the same as the one that flowed through the
ancient landscape over one thousand years ago, it has meandered extensively into what it is
today. There are numerous archaeological sites in Chaco Canyon located within the Chaco Wash
corridor, and the river periodically flows over some of these sites (Allen 2002). However, the
amplified impact of climate change, in the form of extreme weather, combined with the natural
meandering of the Chaco Wash, poses a future threat that could result in a rapid rate of
deterioration for sites directly within the Chaco Wash corridor.
Chaco Canyon is renowned for its remarkable architecture, including massive multistory
stone structures known as “Great Houses.” The most famous of these is Pueblo Bonito, which
has over 600 rooms. The construction of these buildings required sophisticated engineering
techniques, including precise masonry, advanced irrigation systems, and astronomical
alignments. Chaco Canyon was a place of major cultural significance and played a crucial role in

8
the development of Ancestral Puebloan society, serving as a ceremonial, economic, and
administrative center. Evidence suggests that Chaco Canyon was connected to a vast trade
network stretching across the American Southwest and even into Mesoamerica. Goods such as
turquoise, shell, and macaw feathers were traded over long distances. There are several structures
inside the canyon used for religious ceremonies and astronomical observations, with many of
these buildings aligned with solstices and equinoxes. Despite its significance, Chaco Canyon was
mysteriously abandoned around the 12th century, with the reasons still debated among
archaeologists. Chaco Canyon continues to be a focal point for archaeological research, with
ongoing excavations and studies shedding new light on Ancestral Puebloan culture and its
significance in pre-Columbian North America.
1.3.2 Site Bc 53 Background
During the initial stages of this research, Bc 53 was thought to be a completely
unidentified structure because, in the majority of the maps of Chaco Canyon that define the
location of all the known ruins, Bc 53 is not identified. The Chaco Research Archive is one of
the only sources that correctly identifies the site on their map of the historical ruins of Chaco
Canyon (Chaco Research Archive 2010). In the literature, site Bc 53 is not mapped and is only
mentioned a handful of times by other archaeological investigations but not in explicit detail
(Watson 2012). Why was this site not correctly identified in the literature? Why is the site so
hard to identify, even when using high-resolution satellite imagery? Why are there no modern
walking trails surrounding the ruin as all the other ruins have in the area? After some further
investigation, some interesting findings transpired. Site Bc 53 was excavated in 1941 by a team
directed by Frank H. H. Roberts Jr.; however, the results of the excavation never made it into the
published literature (Ditto 2017), leading to no other academic source incorporating the structure

9
on their maps (Chaco Research Archive 2010). In addition, this off-the-record excavation was
conducted by students at the University of New Mexico, and the School of American Research
field schools were attending a field school, which is where archaeologists first get trained to do
field work under the direction of a head archaeologist. Instead of conducting the archaeological
survey, documenting the findings, and publishing it in the literature, the only records of the
research are confined to photos and field notes. Because Frank H.H. Roberts worked at the
Smithsonian as the Director of the Smithsonian Institution River Basin Surveys, the photos, field
notes, and seldom artifacts are sitting in the Smithsonian Archives; however, they are fortunately
digitized as well (Rappaport 2011). This research aims to officially introduce site Bc 53 into the
literature and define the exact location and shape of the ancient structure (Friedman, Sofaer, and
Weiner 2021).
There are three main ancient staircases at Chaco Canyon; the largest is called Jackson’s
Staircase and overlooks the Chetro Ketl ruin. The second staircase, which is referred to as the HP
Staircase during this project, is located behind the Hungo Pavi ruin. The final main staircase at
Chaco Canyon sits 77 meters southwest of site Bc 53. This ancient staircase built into the
cliffside is one of the focuses of the LCP analysis conducted for this research, which will be
further referred to as the FP Staircase for the original excavators of the nearby site Bc 53 Frank
H. H. Roberts Jr. and Paul Reiter. An LCP analysis on site Bc 53 and the nearby FP Chacoan
Staircase is conducted to investigate a possible connection between the site and the staircase with
the goal of better understanding the Ancestral Puebloan roadway network and how the Anasazi
incorporated their roadways and architecture into the surrounding landscape.

10
1.3.3 Anasazi Culture and Society
Chaco Canyon is not an easy place to center a civilization, with temperatures in the
summer reaching over 100 degrees and below 20 degrees in the winter, especially without the
conveniences of modern technology. Despite the challenging climate, the Anasazi culture at
Chaco Canyon between 850 CE and 1200 CE was a lively community with ancient roadways
that were used to enhance trade and travel as well as irrigation systems for their crops (Judge
1988).
The Anasazi built their enormous and elaborate Great Houses throughout the canyon,
with the largest one, Pueblo Bonito, containing over 600 rooms. Different masonry styles of
these Great Houses can be dated, reflecting the development of Chacoan architecture. In addition
to the Great Houses, the Anasazi also built structures called Kivas. Kivas are round semisubterranean rooms central to Chacoan architecture and likely served ceremonial and social
functions but did not serve as living quarters. These Kivas evolved from earlier pit structures and
varied in size and purpose (Judge 1988).
The Anasazi are famous for their ceramic works. Ceramics became a staple over
basketware mainly due to their cultural shift away from hunting and toward agriculture, leading
to a more sedentary way of life. Ceramics offer several major advantages over basket ware,
including that they take less time to construct, are watertight, can be placed directly over fire,
and do not age. The style and design of ceramics have evolved heavily over time. Ceramic types
can be used to establish temporal sequences, aiding archaeologists in dating sites and
understanding cultural changes. Ceramics played a significant role in the lives of the Anasazi,
evolving from plain grayware to intricately decorated vessels over their period of inhabiting
Chaco Canyon.

11
The Anasazi used various tools and objects in their daily lives, including hunting, leisure
activities, clothing making, and food storage. Until 700 CE, they relied heavily on hunting big
game such as elk, mule deer, mountain sheep, and bison. Then, they transitioned to a more
settled farming lifestyle, cultivating crops like corn, beans, and squash. Additionally, the
inhabitants of Chaco Canyon practiced dry farming and utilized canals and ditches to collect
runoff during summer storms.
The inhabitants of Chaco Canyon were also skilled artisans, producing a range of items
such as baskets, ceramics, stone tools, jewelry, and ornaments. The Anasazi valued color greatly.
Apart from the colorful stones used for jewelry, various minerals were ground into pigments for
paints. Hematite (red), limonite (yellow), azurite (blue), malachite (green), and gypsum (white)
were ground on stone mortars (NPS 2003). These minerals were then mixed with water or
vegetable grease and used to decorate various objects, including wooden items such as arrows.
Pigments were also employed to paint murals on plastered walls. Unfortunately, only a few of
these murals have survived. The Anasazi engaged in both regional and long-distance trade,
importing and exporting goods like ceramics, cherts for tools and weapons, as well as exotic
items such as shells from the Gulf of California, turquoise from Cerrillos, New Mexico, and live
macaws most likely originating from Mesoamerica (NPS 2003).
Even though the Anasazi lacked formal written language, they utilized rock art, including
petroglyphs and pictographs, for communication. Petroglyphs are designs carved or pecked into
the rock, and pictographs are designs just painted on the rock, which in contrast, erode quickly
over time. These images likely conveyed group/familial affiliation, historical events, and
ceremonial or ritual information. Other rock art at Chaco Canyon depicts important events during
migrations, songs, stories, and some for the sake of artistic expression.

12
1.3.4 The Ancient Chacoan Roadway Network
The Chacoan road network is one of the most remarkable and enduring legacies of the
Anasazi. These roads, which often span significant distances across the American Southwest,
were more than just pathways; they were cultural and ceremonial symbols that connected various
communities across the region. The roads connected Chaco Canyon to outlying communities,
known as "Chacoan outliers" (Snead 2012). These roads also provided access to resources such
as water, timber, and building materials, which were crucial for sustaining the Anasazi
population and their construction projects. They facilitated the movement of people, goods, and
ceremonial pilgrimages, which, in turn, strengthened the social, economic, and political ties
within the Ancestral Puebloan world.
The Anasazi constructed this extensive network of roads that radiated outward from the
central hub of Chaco Canyon. The roadways were engineered with precision and intent, often
constructed in straight lines over long distances despite the challenging terrain. The roads were
vast, between 8 and 12 meters wide, and were often flanked by masonry walls, earthen berms,
and edging stones. These roadways also had solid foundations, with the depth of these roads
ranging from 10 cm to 50 cm (Vivian 1997a). The construction of these roads would have
required substantial labor and resources, indicating their importance to the Anasazi. They were
built using various methods, including clearing vegetation, leveling the ground, and creating the
foundation for the roads. The road surfaces were often compacted, with some evidence
suggesting the use of gravel or other materials to stabilize them. In some areas, the roads were
elevated, possibly to prevent waterlogging or erosion, and there are even instances of stairways
and ramps being constructed to help the roads traverse steep terrain.
It is difficult to determine the exact timeframe for constructing these roads. However,
taking into account the scale and complexity of the network, it is likely that they were built over

13
several generations, possibly spanning the entire occupation of the Anasazi in Chaco Canyon
from the 9th to the 12th centuries. The roadways are thought to have assisted with maintaining
order within Anasazi society (Vivian 1997b). The roadways' longevity and continued use
throughout the Anasazi inhabitance of Chaco Canyon emphasize their significance in the
Chacoan world, both as physical connectors and as symbolic representations of the Anasazi
cultural and spiritual life.
Beyond their practical use for movement and transportation, the roadways served as an
integral part of the Ancestral Puebloan worldview, which holds a deep connection between the
landscape and astronomy. The roads served as ceremonial routes and, at times, represented the
connection between the living with the spiritual world and the past with the present. The
roadways are also thought to have served as a symbolic meaning of relationships in Anasazi
society; a solid and clear path represents a stable relationship between groups. At the same time,
unkept roads may symbolize a fading connection (Snead 2012). These roads were meticulously
planned and constructed, serving as one of the many examples of the engineering prowess and
communal efforts of the Ancestral Pueblo.
As one of the descending tribes of the Anasazi, the Apache inherited and adapted aspects
of Anasazi traditions, maintaining a deep connection to the land and its significance. The Apache
see the landscape as a living entity, where every mountain, river, and landmark held a story and
symbolized a piece of their cultural heritage. For the ancient Apache, their movement through
the landscape was not merely a matter of efficient travel; these embedded symbolic meanings
also influenced their movement. For instance, the ancient Apache often navigated their
surroundings based on oral histories and ancestral paths, reflecting the belief that places were not
just physical locations but also carriers of cultural memory and spiritual significance. This

14
connection is evident in how places are not just physical locations but are woven into the fabric
of Apache identity and spirituality, most likely inherited from their ancestors (Basso 1996). Even
for the ancestral Apache, their navigation patterns were influenced by the belief that moving
through these places allowed them to honor their ancestors (the Anasazi) and maintain a
connection to their heritage.
Modern-day Apache communities continue to have a deep connection and reverence for
their ancestral landscape. This bond with the land is a significant aspect of their cultural and
spiritual identity. Many modern-day Apache people still hold traditional knowledge and practices
related to their environment, viewing the landscape as sacred and integral to their heritage.
Traditional stories, ceremonies, and historical practices are often tied to specific places,
reinforcing their significance. The Apache's respect for the landscape is also reflected in their
efforts to protect and preserve sacred sites, natural resources, and cultural practices that have
been passed down through generations.
1.3.5 Descendants of the Anasazi and Chaco Canyon’s Role in Modern Culture
Toward the end of the Anasazi residency in Chaco Canyon during the 1100s and 1200s,
changes took place in Chaco as new construction in the area reduced and its position as a
regional center in the American Southwest. However, Chaco's influence persisted in other
regional centers to the north, south, and west, such as Aztec, Mesa Verde, and the Chuska
Mountains. As time passed, people gradually moved away from the traditional Chacoan ways of
life and ventured into new areas across the Southwest. This migration out of the canyon led to
interactions with other groups of people, leading to the transformation of Anasazi culture into the
Native American tribes that are seen in the American Southwest today.

15
The descendants of the Anasazi people of Chaco Canyon almost 1000 years ago are
members of the 20 Native American tribes of New Mexico, Utah, Colorado, and Arizona. For
the Modern-day Pueblo tribes of New Mexico, the Hopi of Arizona, the Navajo, the Southern
Utes, Mountain Utes, Ysleta del Sur Pueblo of Texas, Zuni Tribe, and Jicarilla Apache Nation,
Mescalero Apache Tribe, the history of their ancestors, their accomplishments, and traditions are
passed down to future generations to maintain the connection to their ancestors. (NPS 2003)
Chaco Canyon holds significant cultural and spiritual importance for many Southwestern Native
Americans today, as it is a central location along their ancestors' sacred migration paths. It is
regarded as a place of reverence and great significance that deserves to be treated with respect
(NPS 2024). According to the UNESCO World Heritage Convention, there is evidence that there
are ancient roadways and peripheral communities, also with Great Houses, located beyond the
current property boundary of Chaco Canyon National Historical Park. However, these outlier
roadways and communities were not taken into account during the UNESCO inscription process.
Since Chaco Canyon was inscribed on the UNESCO World Heritage List, steps have been taken
to slow down its rate of deterioration, such as partial site reburial, defined fencing, and increased
park ranger patrolling. Even though there are no current negative impacts directly on the
property, there has been an increase in the potential hazards posed by the development of
surrounding areas, which includes utilities, roads, energy exploration, extraction, as well as mass
transportation projects (UNESCO 2024). Additionally, a law was passed in 2023 that approved a
10-mile buffer zone around Chaco Canyon National Historical Park, meaning that it is no longer
open to development. The buffer zone surrounding Chaco Canyon has become a point of
contention among various law and policymakers, including tribal councils, the State of New
Mexico, the National Parks Service, and the Department of the Interior. Some decision-makers

16
advocate for the buffer zone in order to protect Chaco Canyon as much as possible. However,
other stakeholders, such as tribal members from the Navajo Nation, who currently inhabit the
area and were allotted these parcels of land many generations ago, do not view this buffer zone
as fair. This is because it essentially freezes any land they own inside the 10-mile buffer zone
surrounding Chaco Canyon. While the newly passed legislation only controls the federal land
inside the buffer zone, and privately owned land is still free to allow resource extraction for
profit, the demand has dropped significantly. The oil and gas industry has pulled back its
investment into this area due to the new policy, which leaves many tribal members holding land
that they have to repurpose or sell now that it will no longer be a reliable revenue stream
(Adomaitis 2023).
Despite its importance, Chaco Canyon faces several contemporary threats that jeopardize
its preservation and integrity. The region is susceptible to threats from resource extraction
activities, including drilling and gas production. These activities can lead to habitat destruction,
landscape alteration, and contamination of archaeological sites, posing a significant risk to the
cultural and environmental heritage of Chaco Canyon. According to the National Parks
Conservation Association, the largest methane hotspot, covering 2500 square miles, is located
above Chaco Canyon. This hotspot is a direct result of drilling and gas production in the area.
More than 75% of the residents of San Juan County, where Chaco Canyon is located, live within
half a mile of gas and oil infrastructure. Additionally, over 91% of the government land
surrounding Chaco Canyon National Historical Park is leased to the oil and gas industry by the
Bureau of Land Management (NPCA 2023). Additionally, the fragile sandstone formations in
Chaco Canyon are susceptible to erosion, which can be exacerbated by natural processes such as
wind and water as well as human activities. The effects of climate change, including increased

17
temperatures, altered precipitation patterns, and more frequent extreme weather, could impact the
fragile desert ecosystem of Chaco Canyon and its archaeological resources. Illegal looting of
archaeological sites for artifacts and vandalism of ancient structures also continues to be
significant threats to the preservation of Chaco Canyon’s cultural heritage. There is also an
important challenge posed to managing and protecting the cultural resources of Chaco Canyon,
including archaeological sites and artifacts, which requires significant resources and coordination
among various stakeholders, including tribal communities, government agencies, and
conservation organizations.
1.4 Project Overview
This thesis research project has three main objectives. First, a supervised image
classification is conducted on satellite imagery of Chaco Canyon, with the aim of identifying
new archaeological areas of interest based on the spectral signatures of the known ruins and
roadways. Second, LiDAR data is used to investigate a concealed site, and then a LCP analysis is
performed to model potential travel routes to nearby ruins from the concealed site. For the final
objective, a 3D visualization of Chaco Canyon is generated using LiDAR point cloud data and
high-resolution satellite imagery. The overarching goal of this research is to support ongoing
cultural heritage preservation efforts at Chaco Canyon. The findings will benefit the National
Parks Service, associated tribes, conservation groups, and the broader academic and public
communities by providing a complete 3D visualization that can be utilized for educational
purposes, preservation efforts, informing public policy, and as a foundation for future
archaeological research at Chaco Canyon. The findings of this research also inform current
efforts to protect this archaeological gem from the encroaching oil drilling and gas production in

18
the area. The knowledge gained from this project not only enriches our understanding of the past
but also helps preserve and protect valuable archaeological sites.
There are three primary methodologies for this research project. First, a supervised image
classification is conducted utilizing both the random trees and the support vector machine
algorithms. For the second methodology, site Bc 53 is further investigated, and a LCP analysis is
conducted, with a focus on site Bc 53 and its neighboring sites. For the final primary
methodology, a contiguous 3D visualization of Chaco Canyon will be generated, incorporating
high-resolution satellite imagery, the LiDAR-derived DEM, and custom symbology to identify
the major ruins and ancient roadways.
This research is based on two main pieces of data: the 30 cm and 1 m resolution satellite
images of Chaco Canyon acquired from Land Info Worldwide Mapping and the open-source
LiDAR data with a spatial resolution of 1 m acquired from OpenTopography. One of the primary
goals of this research is to digitally protect the archaeological ruins at Chaco Canyon. "Digitally
protect" refers to capturing and preserving the current state of these sites using advanced remote
sensing technologies before further degradation of the archaeological sites occurs. Specifically
for this research project, a contiguous, realistic 3D model of Chaco Canyon and its
archaeological sites is created at a spatial resolution of 1-meter, utilizing high-resolution satellite
imagery and a LiDAR-derived elevation model. This 3D model ensures a detailed and accurate
representation of the landscape, structures, and ancient roadways. By documenting these sites in
a digital format, the current condition of the sites can be safeguarded for future generations. This
3D model "digitally protects" the archaeological sites, allowing for continued study and
appreciation of the ruins, even if they are no longer physically visible due to natural deterioration
as well as human-caused factors.

19
1.5 Document Overview
The manuscript begins with an introductory chapter that covers the motivations for this
thesis research project and provides background information on the study areas, Chaco Canyon
National Historical Park, and Site Bc 53. The second chapter lays the foundation for this research
by outlining related work and previous research. It discusses the related literature for the three
main analyses: the supervised image classification of Chaco Canyon, the examination of site Bc
53 and its LCP analyses, and the creation of a 3D visualization of Chaco Canyon. The third
chapter of this manuscript outlines the research design and methodology for this project. It starts
with a brief overview of the overall methodology of the project. The following section delves
into the data necessary to complete this research. The chapter concludes with a final section that
details the specific methodologies for each of the three main analyses. The fourth chapter
examines the results and products created from this research project. It analyzes the results of the
supervised image classification, the LCP analysis, and the physical condition and preservation
state of site Bc 53. The fifth and final chapter focuses on the conclusions that can be drawn from
this research and the objective discussion of the research and its process. It reviews the results,
discusses their implications, and explores the limitations of the project and future directions for
related research.

20
Chapter 2 Related Work
In this chapter, the existing literature and previous research on supervised image classification of
high-resolution satellite imagery is discussed. Subsequently, the relevant literature on Site Bc 53,
as well as LCP analyses in similar archaeological contexts, is explored. This exploration is
followed by a review of previous work and research on the creation of 3D models and
visualizations of archaeological sites at Chaco Canyon. Additionally, previous research and
literature on Anasazi culture and society is examined. Finally, the descendants of the Anasazi
today and the role of Chaco Canyon in the modern world are considered.
2.1 Supervised Image Classification
This section discusses the use of supervised image classification in related archaeological
research. A previous study conducted by Budge in 1981 is first examined, which employed
supervised image classification techniques at Chaco Canyon to map the ancient roadway network
and to define the location of the major ruins. Then, a more recent research study conducted by
Witts, which also employed a supervised image classification on archaeological sites near Chaco
Canyon, is evaluated. Additionally, a research study conducted by Keeney that focused on
utilizing a supervised image classification on high-resolution imagery in Alaska to locate new
archaeological sites is also reviewed. The limitations of these studies and opportunities for
improvement in the use of remote sensing techniques are highlighted. This section also discusses
how this thesis addresses the existing gap in remote sensing techniques applied to Chaco
Canyon. It introduces a modernized approach that utilizes high-quality multi-spectral imagery,
sub-meter spatial resolution, and the advanced capabilities of the Esri platform ArcGIS Pro for
completing supervised image classification.

21
2.1.1 Previous Archaeological Research Employing Image Classification
Previous research conducted at Chaco Canyon has employed an image classification
analysis, researchers such as Budge and their team had success using this type of analysis to
locate ancient structures and roadways and define the agricultural landscape at Chaco Canyon
(Budge 1981). A limitation to much of the research that has been conducted at Chaco Canyon
employing a supervised image classification analysis is that they are dated and have not been
revisited. While Budge does provide steps for analysis techniques, the workflow is from the 80s
and is based on aerial imagery captured in the late 70s, which does not come close to the spatial
resolution that aerial imagery today can offer.
In terms of more recent research, Witt's research study was conducted in 2010 on an area
near Chaco Canyon called Farmington, New Mexico, which was seeking similar results as this
thesis project but was unable to identify any new ancient roadways or the locations of any
potential archaeological sites (Witts 2010). One area where Witts' research is lacking is that it is
based on very zoomed-out imagery at both 30 m and 90 m resolution to try and encompass the
whole Middle San Juan Region. Due to the low spatial resolution, the image classification that
Witts conducted did not produce any usable results.
A previous research study conducted by Keeney and their team examined a site in Alaska
at a more zoomed-in scale in contrast to Witts' research. Keeney's research study aims to test
whether using satellite imagery and other remote sensing technology can help locate
archaeological sites associated with mobile hunter-gatherer groups in Alaska's Brooks Range.
Using 1-meter resolution IKONOS imagery, Keeney and the team employed both unsupervised
and supervised image classifications, revealing a spectral phenomenon associated with
archaeological sites, particularly those found under dense willow stands. The Keeney study
suggests that supervised and unsupervised image classification methods could help improve

22
survey strategies and the identification of archaeological sites when operating under the right
circumstances, such as working with aerial imagery with a high enough spatial resolution to
identify the nuances in the landscape (Keeney 2015).
2.1.2 Foundational Literature on Supervised Image Classification in Archaeology
To carry out the supervised image classification for this thesis project, several key
literature sources are utilized to guide the process. In this research, a methodology outlined by
Keeney in 2015 that uses supervised and unsupervised image classification techniques on highresolution satellite imagery to locate archaeological features in a cost-effective and efficient
manner is referenced (Keeney 2015). This source is a research project conducted by Argyrou and
their team in 2023 that highlights the advantages of using supervised image classification on
high-resolution aerial imagery in archaeological research. This research also provides
background information on previous work done using remote sensing tools to identify areas of
archaeological significance (Argyrou et al. 2023). The next source that is utilized for this
research is a traditional archaeological research study completed by Gumerman in 1977, which
employs remote sensing techniques and aerial imagery at Chaco Canyon National Historical Park
to map cultural features such as the ancient roadways and determine the extent and
characteristics of the roadways that converge on the Pueblo Alto ruin within Chaco Canyon
(Gumerman 1977). Finally, the fourth key literature source is a research study conducted by
Witts in 2010, which focuses on the application of remote sensing techniques on similar
archaeological sites created by the Ancestral Puebloans.
2.2 Site Bc 53 and Least Cost Path Analysis
This section discusses the previous research and literature gap on site Bc 53, and the
previous LCP analyses conducted at Chaco Canyon. The existing literature on site Bc 53 is

23
minimal and inconclusive due to the unsuccessful excavation that essentially destroyed the site.
However, by using sources such as the Chaco Research Archive, this thesis project aims to
reidentify and map site Bc 53. Additionally, this section presents previous research studies that
utilized LCP analysis, such as Alvez's study on the Beaker culture in Northeast England during
the Bronze Age and Field's research on Chaco Canyon, which explores the ancient Chacoan
roadway network.
2.2.1 Previous Research Regarding Site Bc 53
In terms of the previous research that has discussed site Bc 53, it is minimal and
primarily based on only field notes and photos of the site in the Smithsonian Online Virtual
Archive (Rappaport 2011). Due to the standard archaeological practices of the time during the
initial excavation of site Bc 53, which did not prioritize site preservation for future research, the
excavation was somewhat destructive. Additionally, the tools and techniques available in the
1940s lacked the precision of modern tools and methodologies, leading to potential inaccuracies.
The excavation was also part of an archaeological field school, which differs from the work of a
fully trained archaeological team. The absence of published findings from this excavation has
further contributed to the limited scholarly understanding of site Bc 53.
The literature following this excavation only briefly mentions site Bc 53 as being
somewhere near the Casa Rinconada community. However, it is never discussed in great detail,
and it is seldom correctly identified on the maps in the literature (Ditto 2017). The
inconsistencies in the literature and lack of overall information regarding site Bc 53 are the direct
result of the failure to properly document the site during this excavation that can never be
undone. The majority of the information about site Bc 53 was discovered for this project by
using the website Chaco Research Archive, that does an excellent job of mapping and locating

24
site Bc 53, which most of the current literature on Chaco Canyon fails to do (Chaco Research
Archive 2010).
Because one of the main goals of this research project is to “digitally protect” the ruins at
Chaco Canyon using LiDAR technology to generate a hyper-realistic 3D model, previous
research conducted by Guo in 2019 is examined. The research study conducted by Guo was
fascinating because it developed and incorporated an algorithm that allows for quick and
accurate registration of multiple LiDAR point cloud sites. During this research project, the
presented approach employing the newly created algorithm was successfully tested at the burial
site in Yaoheyuan, Ningxia, and has been shown to surpass mainstream open-source algorithms
in terms of accuracy and efficiency. For this research project an orthophoto generation system
was created that automatically generates high precision orthophoto maps, which makes
comprehensive digitalization of archaeological sites possible (Guo 2022).
A research study completed by Alvez, focused on the social and economic trends of the
Beaker culture in Northeast England during the Bronze Age, using a LCP analysis. The LCP
conducted by provided in this research very informative results about the ancient Beaker
civilization and their mobility routes. By generating the LCPs that take into account the local
topography, this research study was able to model spatial connections between different
archaeological sites. The results of Alvez's study show paths that mostly coincide with local
variations in the terrain and follow watercourses. This approach has the potential to help
archaeologists understand how humans moved during the Bronze Age (Alvez 2016). The main
limitation of Alvez's research is that it only utilizes the general topography as the main driving
factor for the LCP analysis.

25
Another research study conducted by Field in 2023 introduces a systematic approach for
creating, visualizing, and comparing individual Chaco roadway profiles using LiDAR-derived
elevation data. The method presented in this research study utilizes elevation values obtained
from LiDAR-derived digital elevation models. The goal of Field’s research is to establish a
common form of ground-truthing on the Chaco roadways and evaluate its frequency across nonground-truthed roadways. This approach provides a valuable tool for documenting and
comparing ancient roadways using remotely sensed data; it outlines what anomalies to look for
in the LiDAR that may be ancient roadways and ruins, especially in landscapes where ground
truthing is difficult (Field 2023).
2.2.2 Critical Literature for Site Bc 53 and LCP Analysis
One of the main goals of this thesis research is conduct a comprehensive LCP analysis
and investigation of site Bc 53 by drawing upon critical literature encompassing digital archives,
LiDAR-based roadway profiling methodologies, 3D visualization techniques, and previous LCP
analyses in similar archaeological contexts. The Smithsonian Online Virtual Archives are first
used to access the notes, photos, and artifacts list for site Bc 53 to further strengthen the overall
background of site Bc 53 (Rappaport 2011). The Chaco Research Archive is also utilized as a
source for information regarding site Bc 53 (Chaco Research Archive 2010). The techniques
from Field's 2023 research study that introduces a systematic approach for creating, visualizing,
and comparing Chaco roadway profiles using LiDAR-derived elevation data is also incorporated
into this thesis research to identify the ancient roadways. The methods presented in Field's study
utilize elevation values obtained from LiDAR-derived digital elevation models, with the goal of
establishing the common form of ground-truthed Chaco roadways and evaluating its frequency
across non-ground-truthed roadways. This approach provides a valuable tool for documenting

26
and comparing ancient roadways using remotely sensed data and knowing what anomalies to
look for in the LiDAR data, especially in landscapes where ground truthing is difficult (Field
2023).
The 2019 research also conducted by Field is utilized throughout this thesis research
project. The study conducted by Field focused on modeling the ancient Chaco Canyon roadway
network using a large-scale LCP analysis. Some of the methodologies found in Field's study are
adapted to be better integrated into this smaller-scale and more localized LCP analysis for site Bc
53 and the surrounding ruins (Field 2019).
Another key piece of research to this thesis project is the study conducted by Guo in
2022. Guo's research is informative about how LiDAR point cloud data can be utilized to protect
archaeological sites digitally, which is very influential to this research (Guo 2022). The research
conducted by Guo serves as an essential piece of literature that sheds light on how LiDAR data
can be optimally leveraged to “digitally protect” the Chaco Canyon ruins.
Alvez's research project, which employed an LCP analysis on the Beaker culture of
Northeast England during the Bronze Age, is the final critical piece of literature to guide this
section of the thesis research. The overall goal of Alvez's research is to examine and complete
the LCP analysis between significant locations in the Beaker culture and then examine the results
through the lenses of social and economic trends in ancient Beaker society. Alvez's research is
significant to this thesis project because it clearly outlines how to conduct an LCP analysis in
ArcGIS Pro; the clear instructions that Alvez’s provides are particularly useful when carrying out
the LCP analysis on site Bc 53 for this project. They serve as an initial template for designing a
customized LCP analysis tailored to the specific landscape (Alvez 2016).

27
2.3 3D Visualization of Chaco Canyon
In this section, the use of 3D modeling and LiDAR in fundamental archaeological
research is explored. The aim is to create a 3D visualization of Chaco Canyon by utilizing key
pieces of literature that outline comprehensive workflows and successful applications of LiDAR
technology in archaeological research. Specifically, research studies conducted by Katsianis,
Corns and Shaw, and Carter are examined to provide valuable insights about the application of
context-based systems, helicopter-mounted FLI-MAP 400 LiDAR systems, and ALS technology
in generating detailed 3D models of archaeological sites. These studies are crucial in guiding the
methodology for creating a 3D visualization of Chaco Canyon in this thesis research project.
2.3.1 Visualizing Chaco Canyon and Similar Archaeological Sites in 3D
In terms of the previous research that has been conducted at Chaco Canyon in regard to
generating a 3D model from a LiDAR point cloud, researchers, such as Toeppen, have had
success generating 3D models of individual sites or a small grouping of sites (Toeppen 2022).
One limitation of these individual models is that it is hard to visually understand how these
different archaeological sites fit together spatially to create Chaco Canyon as a whole.
Additionally, when the model is only limited to individual sites, it can be difficult to visualize the
completeness of the Chaco roadway network.
Previous research studies, such as Wills’ 2017 research, have succeeded in generating a
complete 3D model of Chaco Canyon, which was subsequently used to conduct a watershed
analysis on Chaco Canyon (Wills 2017). The limitation of Wills’ research is that it is a blank 3D
model that is not intended to be a visual for the reader; it is explicitly incorporated for
conducting the watershed analysis.

28
Other researchers, such as Carter in 2014, have had success creating 3D visualizations of
archaeological sites in Mesoamerica and their surrounding areas in their entirety, employing
high-resolution satellite imagery to enhance the visualization (Carter 2014). The only limitation
of Carter’s research is that the extent of this project is much more zoomed out than what is
necessary for this thesis project. The research for this project does, however, incorporate
elements from the archaeological research conducted by Carter to generate the 3D visualization
of Chaco Canyon for this research project.
A research study based out of Ireland had much success with the helicopter-mounted
FLI-MAP 400 LiDAR system, which provides highly detailed 3D models of archaeological sites
(Corns and Shaw 2009). The main drawback to this research is the amount of funding that needs
to be acquired to conduct this type of research project at such high quality, as well as the use of
occupied aircraft to capture the LiDAR data.
2.3.2 Fundamental Archaeological Research Employing 3D Modeling and LiDAR
There are several key pieces of literature that are utilized when creating the 3D
visualization of Chaco Canyon. A research study conducted by Katsianis in 2008 explores the
application of a context-based system for recording the excavation process across various
archaeological projects in Northern Greece. The focus of Katsianis' research is to present a
formal data model and a complete digital workflow for documenting the excavation process in
3D at the prehistoric site of Paliambela Kolindros in Greece. The workflow from Katsianis’
research can be applied to other archaeological sites to create a 3D model using ArcGIS Pro and
other Esri products (Katsianis 2008). Katsianis' research is significant to this thesis project
because it goes in-depth and outlines a comprehensive plan about how to create a 3D

29
visualization for an archaeological site using ArcGIS Pro, which can be directly applied to the
3D modeling conducted for this research project.
Another essential piece of literature that is used for this thesis research is a study
conducted by Corns and Shaw in 2009, which focused on the Discovery Programme Centre for
Archaeology and Innovation in Ireland and how the Programme Centre transitioned from
terrestrial-based surveys to digital aerial stereo photogrammetry and then finally adopted a
helicopter-mounted FLI-MAP 400 LiDAR system, which provides highly detailed 3D models of
archaeological sites. Corns and Shaw's research elaborates on successful applications at sites like
Newtown Jerpoint and the Hill of Tara, discussing the efficacy of the new system in data
processing, vegetation removal, and field inspections. During this research, the newly generated
new methodology is evaluated against the more traditional LiDAR and ground-based approaches
(Corns and Shaw 2009). The research conducted by Corns and Shaw is relevant to this thesis
research because it provides a professional workflow for creating 3D models of archaeological
sites using LiDAR. While this thesis project does not have anywhere near the same funding or
equipment as Corns and Shaw's research, it incorporates bits and pieces of the methodologies
described by the Discovery Programme Centre for Archaeology and Innovation for generating
3D models of archaeological sites.
The groundbreaking study conducted by Carter in 2014 illuminates the pivotal role of
airborne mapping LiDAR technology in archaeological research, particularly in densely
vegetated regions like Mesoamerica, signifying a revolutionary shift in the field of archaeology.
Carter's research provides a comprehensive understanding of ALS technology, its collection
process, and its implications for future research. This research further emphasizes how ALS
observations and data are highly customizable to fit the needs of the research project (Carter

30
2014). Carter's research is essential to this thesis research as it provides valuable information
about the significance of remote sensing tools in the field of archaeology. It also sheds light on
other similar studies that have employed LiDAR with the ultimate goal of locating obstructed
archaeological sites.

31
Chapter 3 Research Design and Methodology
In this chapter, an overview of the research design and methodology for this project is first
provided. The following section in this chapter examines the data utilized to complete this
research. The chapter concludes by outlining the methodologies for the three main analyses in
this research project: the exploratory analysis of Chaco Canyon, the site Bc 53 analysis, and the
generation of the 3D visualization of Chaco Canyon
3.1 Overview
This section outlines the methodology for the following three processes: supervised
image classification, the LCP analysis and assessment of site Bc 53, and the generation of the 3D
visualization of Chaco Canyon are outlined. These methodologies are separated into three
separate workflows, which are further detailed in this section. The first subsection discusses the
initial exploration of Chaco Canyon for this project, the process for both the RT and SVM
supervised image classifications, and specific investigation into Site Bc 53. The following
subsection discusses the workflow for examining the physical condition of Site Bc 53 using
remote sensing technologies, as well as how to generate the LCP analysis for the site. This
methodology section concludes by explaining the workflow for generating the final 3D
visualization of Chaco Canyon.
3.2 Data
This research is based on two main sources of data: high-resolution satellite imagery of
Chaco Canyon obtained from Land Info Worldwide Mapping and the open-source LiDAR data
package acquired from OpenTopography, as displayed in Table 1. Google Earth satellite
imagery, captured by Airbus in 2024 at a spatial resolution of 15 cm, is initially used to explore

32
the Chaco Canyon area as displayed in Figure 3. However, this imagery cannot be exported
outside the Google Earth Engine and is only available as a Tile Layer in ArcGIS Pro through
ESRI’s Living Atlas, which cannot be used for the analyses in this research.
Figure 3. Google Earth view of Chaco Canyon
Alternatively, to complete this research, satellite imagery of the study area at 30 cm and 1
m resolution in both multispectral and panchromatic formats is obtained and is comparable to
what Google Earth provides. The multispectral imagery captures all visible light waves as well
as near-infrared wavelengths as displayed in Figure 4. This imagery is acquired from Land Info
Worldwide, which is based in Denver, Colorado, and has access to the Pléiades Satellite
Constellation and is displayed in. Once the imagery is obtained, the imagery is orthorectified to
ensure that it is accurately displayed in the correct location in ArcGIS Pro.

33
Figure 4. Raw Multispectral Satellite Imagery
The other key data utilized in the research is a LiDAR point cloud and a Digital
Elevation Model (DEM) of the study area, which allows for the creation of the most accurate 3D
visualization of Chaco Canyon possible as shown in Figure 5. This open-source data package,
which includes the LiDAR point cloud data and the DEM derived from the LiDAR, is found
available on the website OpenTopography. This data was collected from a previous research
study focused on an ancient watershed analysis of Chaco Canyon (Dorshow 2010). Both the
LiDAR point cloud and the DEM have a spatial resolution of 1 m, which pairs nicely with the 1
m imagery obtained from Land Info to create a realistic 3D visualization of Chaco Canyon.

34
Figure 5. Raw LiDAR-derived DEM
Table 1. Data
Name Content Format Attributes Precision Availability Price
Pleiades
Satellite
Imagery
Satellite
Imagery
GeoTIFF 16-bit pan + 6 band
multispectral
bundle
Panchromatic/
Multispectral:
30cm;
Multispectral:
1m resolution
Private
Access;
Land Info
Worldwide
Mapping
$285
Chaco
Canyon,
NM:
Simulating
Dynamic
Hydrological
Processes
LiDAR
Point
Cloud
and
DEM
(TIN)
.LAZ for
point
cloud and
GeoTIFF
for DEM
Total number of
points: 298,677,157
pts
Point density: 4.46
pts/m2
1m spatial
resolution
Open
Source;
acquired
from OpenTopography
Free
Google
Earth
Satellite
Imagery
Satellite
Imagery
(Source
: Airbus
Imagery
2024)
Data
Layer in
ArcGIS
Pro
found in
ESRI’s
Living
Atlas
Panchromatic: 65
cm (nadir) to 73 cm
(20⁰ off-nadir);
Multispectral: 2.62
m (nadir) to 2.90 m
(20⁰ off-nadir);
4-band
Multispectral:
(blue, green, red
and near infrared)
15 cm spatial
resolution in
this area on
Google Earth
Open
Source
Free

35
3.3 Exploratory Analysis Methodology
The methodology for this research project begins by first acquiring satellite imagery of
Chaco Canyon National Historical Park. An exploratory analysis is conducted utilizing the highresolution satellite imagery from both Google Earth and Land Info Worldwide Mapping.
The open-source platform Google Earth is first utilized to explore the overall Chaco
Canyon study area. When inside the Google Earth platform and viewing the Chaco Canyon study
area, the ruins are very identifiable, displaying all the nuances in the structures. The highresolution imagery that Google Earth uses was captured by Airbus in 2024 at a remarkable 15 cm
spatial resolution. During this initial exploration phase, the Bc 53 ruin was first identified, which
has a structural outline that is barely visible in comparison to the other nearby ruins, even at a
spatial resolution of 15 cm. The imagery cannot be exported outside Google Earth, and it is only
available as a Tile Layer in ArcGIS Pro through ESRI’s Living Atlas; alternatively, for this
project, the imagery acquired from Land Info Worldwide Mapping at both 30 cm and 1 m spatial
resolutions are utilized.
Once the satellite imagery is acquired from Land Info, it is reviewed to ensure that the
spatial resolution is fine enough to conduct the supervised image classification properly. If the
spatial resolution is too coarse, then the ruins and ancient roadways may not be able to be
identified. The imagery is then brought into ArcGIS Pro, where it is georeferenced to align with
the map in ArcGIS Pro, which has a projected coordinate system set to NAD-83 UTM zone 13.
The next step is to conduct an exploratory analysis of the overall Chaco Canyon study
area by completing a supervised image classification. Running a supervised image classification
will aid in identifying any areas of interest that the human eye cannot detect by using the unique
spectral signatures captured in the multispectral imagery to classify the imagery accurately. The

36
supervised image classification is conducted using 30 cm resolution satellite imagery obtained
from Land Info. Two algorithms are run for this research: random trees and support vector
machine, which both follow the same workflow. For this image classification, the training
samples are broken down into five classes: vegetation, barren, developed, asphalt roads, ancient
roads, and ruins. The training samples are then manually identified in the imagery based on these
classes. Once the training samples are identified, the Classify Raster Tool is utilized to complete
the initial classification of the pixels in the satellite imagery. Following this step, the Reclassify
Raster Tool is then used to reclassify the initial classification according to the training samples.
An accuracy assessment obtained by generating a confusion matrix for both algorithms is used to
determine which algorithm performed the best, classifying the most pixels correctly. The best
algorithm is used to complete the rest of this research.
The findings of the image classification are next used in this research to assist with
definitively identifying the location of Bc 53. The preservation condition and structural shape of
site Bc 53 are also assessed during this step. To conclude this first methodology as displayed in
Figure 6, the classification results are examined to identify the structural outline of the major
ruins at Chaco Canyon, with a specific focus on site Bc 53 to see if the results were able to
distinguish the outline of the structure.

37
Figure 6. Exploratory Analysis Workflow
3.4 Site Bc 53 Analysis Methodology
The next step in the methodology is to zoom further into the study area of Chaco Canyon
and investigate site Bc 53 in more detail. The records and photos pertaining to Bc 53 are first
analyzed to begin exploring the site, which is only available for viewing through the open-source
Smithsonian Online Virtual Archive. During this step, any other related research that discusses
site Bc 53 is examined, with a focus on how it is discussed in the text in relation to the other
nearby ruins.
Once a solid foundational understanding of Bc 53 is established, an LCP analysis is
conducted with Bc 53 as the focus. The LCPs generated from the LCP analyses connect the FP
Chacoan Staircase and Casa Rinconada community with site Bc 53. To begin this LCP analysis,

38
the LiDAR-derived DEM is imported into ArcGIS Pro, and the DEM is first filled using the Fill
Tool. The source point for both of the LCP analyses is site Bc 53. The destination for the first
LCP analysis is site Bc 51, which is on the edge of the Casa Rinconada community of ruins. Site
Bc 51 and the Casa Rinconada community are thought to have been inhabited from the late 11th
century to the early 12th century, coinciding with the suggested occupation of site Bc 53 at the
end of the 11th century (Chaco Research Archive). Site Bc 51 is selected as one of the
destination points for this LCP analysis because it is the closest ruin to site Bc 53. The FP
Chacoan Staircase was also selected as one of the destination points for this LCP analysis
because, while it is not technically a ruin, it is the closest archaeological area of interest in
relation to site Bc 53. Once the source and destination points are defined, the Cost Back Link
Tool is utilized to create a raster that defines the neighbor, which is the next cell on the LCP to
the nearest source. This newly generated back link raster is essential for determining the
direction to move to traverse the LCP from any cell back to the source. The Cost Path Tool is
then used to generate the most optimal path between the source and destination. The same
workflow is repeated to create both LCPs. These LCP results not only indicate the directness of
the most optimal paths, but they also reveal the relationship between the sites based on their
proximity and accessibility to one another. This LCP analysis is critical for understanding how
the Anasazi navigated the canyon landscape by serving as an educated inference of the most
probable routes the Anasazi would have taken to minimize effort expenditure while traveling.
This analysis further sheds light on the daily movement locally within the canyon. Understanding
these LCPs can also provide insights into the functional significance of various sites. For
example, with the FP Chacoan Staircase, by generating these LCPs, it becomes clear that it was a

39
deliberate choice to utilize landscape features to create a rock staircase in the cliffside to connect
the long-distance roadway network with the local roadway network within the canyon.
Additionally, the research conducted on local pathways in Chaco Canyon connecting less
prominent sites, like Bc 53, is minimal. This LCP analysis utilizes empirical spatial data to
identify the most optimal pathways that connect these smaller sites with other nearby
archaeological areas of interest. The findings of this LCP analysis can be used as a guide to
explore the relationships between other smaller, less prominent sites. The results of the LCP
analysis also assist with revealing the true relationship between site Bc 53 and the nearby Casa
Rinconada community of ruins. Site Bc 53 does share the same “Bc” prefix as the nearby ruins
that make up the Casa Rinconada community of structures (e.g., Bc 51 and Bc 50); however, it
appears to be not connected to the community.
To conclude, the physical condition of site Bc 53 is examined using the LiDAR data
acquired from OpenTopography with a spatial resolution of 1 meter. To transform the LiDARderived DEM into a visible product, a multidirectional hillshade layer is generated using the
filled DEM as the input. A copy of this hillshade layer is next generated, and the symbology is
changed to an elevation gradient. By stacking the multidirectional hillshade layer and the newly
created elevation gradient and then adjusting the transparency, the ruins throughout the Chaco
Canyon study area become very distinguishable. The visibility of site Bc 53 is explored and
compared in contrast to the other nearby sites. This second methodology is illustrated in the
Figure 7 workflow diagram.

40
Figure 7. Site Bc 53 Analysis Workflow
3.5 3D Visualization Methodology
The final methodology for this project is to create a 3D visualization of Chaco Canyon
National Historical Park. The multidirectional hillshade layer generated in the previous
methodology is utilized as the base layer of the 3D visualization. The first step is to convert the
2D map with the multidirectional hillshade layer into a Local Scene, which is in three
dimensions. A new 3D scene is generated in the project in ArcGIS Pro. At the bottom of the
contents pane in the new Local Scene, ensure that the LiDAR-derived DEM is being utilized as
the elevation source for the scene.
A 3D model of Chaco Canyon is created using the LiDAR point cloud data with a spatial
resolution of 1 meter. The only visible layer is the multidirectional hillshade layer, which will be

41
used in conjunction with the satellite imagery to generate the most realistic visualization
possible. This 3D model is then wrapped with the 1-meter imagery using the Layer Blend
technique in ArcGIS Pro to mix all of the contributing layers properly. The transparency and
order of the contributing layers are adjusted to achieve the most realistic results.
This step is concluded by finalizing the 3D visualization and incorporating symbology
that defines the location of the ruins as well as the ancient roadways. To ensure that the ruins and
ancient roadway networks are being correctly displayed in the right location, maps that were
created for previous related research are obtained that already have these major ruins and
roadways identified. These reference maps are brought into ArcGIS Pro and georeferenced,
using known points in both the map and the satellite imagery to make the map distances and
locations as accurate as possible. With the georeferenced maps created, line features are created
to define the long-distance and short-distance roadway networks, using the georeferenced map as
a guide. Point features are then created to define the main ruins at Chaco Canyon, using the
georeferenced map as a guide to place the points.
The final generated model is then uploaded to ArcGIS Online for public viewing. The
different layers, such as ancient roadways and structures, can be turned on and off by the viewer,
as displayed in Figure 9. The 3D visualization allows the user to traverse the Chaco Canyon
Landscape, with the ability to tour the major viewpoints digitally at the click of a button. This
third methodology is illustrated in the Figure 8 workflow diagram.

42
Figure 8. 3D Visualization Workflow

43
Figure 9. 3D visualization uploaded to ArcGIS Online

44
Chapter 4 Results
This chapter delves into the specific results of this project. First, this chapter analyzes the results
of the supervised image classification that is conducted on high-resolution satellite imagery. The
next section of this chapter examines the results of the LCP analysis between site BC 53 and its
neighboring sites. The following section investigates the preservation state of Site Bc 53 using
both LiDAR data and the results from the supervised image classification. The chapter then
concludes by exploring the results of the final visualization, which includes the integration of
high-resolution satellite imagery, a LiDAR-derived digital terrain model, and custom symbology.
4.1 Supervised Image Classification
This section analyzes the results from both the random tree and support vector machine
supervised image classification. First, the results of the supervised image classification using
random trees are discussed. The accompanying confusion matrix provides specific accuracy
scores for the image classifications, allowing assessment and comparison of the overall accuracy.
This section concludes by exploring the results of the SVM supervised image classification,
including its confusion matrix, and finally comparing the two classification products.
4.1.1 Random Trees
As part of the exploratory analysis process for this project, a supervised image
classification is conducted on the high-resolution imagery acquired from Land Info Worldwide
Mapping. This high-resolution imagery is exhibited in Figure 10 and Figure 13. Figure 10 is a
snapshot of high-resolution imagery and displays the Casa Rinconada community of ruins as
well as the pathways connecting the sites. The results of the supervised image classification are
depicted in Figure 11, where the random trees algorithm is employed. The color scheme used for

45
this classification has yellow representing barren desert, green for vegetation, red for ruins,
orange for ancient roadways, gray for asphalt, and purple for developed land. Although this
classification misclassified many of the pixels, it still did a decent job of defining and making the
ruins and pathways visible.
By using Figure 10 as a reference, the Casa Rinconada community of ruins and the
connecting pathways can be easily identified in Figure 11. These results also help assess the
condition of site Bc 53, exploring whether it is in a good enough state of preservation to be
identified during the supervised image classification.
Figure 10. Satellite imagery at 30 cm resolution from Pleiades Satellite

46
Figure 11. Classification of Casa Rinconada employing the RT Algorithm
To assess the accuracy of the random trees supervised image classification, a confusion
matrix, also known as an error matrix, is utilized, as displayed in Figure 12. The confusion
matrix assesses the user accuracy, which is also known as Error Type 1 or a False Positive. This
Type 1 Error occurs when a pixel is incorrectly classified as another known class. The matrix
also assesses the producers' accuracy, which is also known as Error Type 2 or a False Negative.
A Type 2 Error occurs when the producers' accuracy does not meet the expectation of the ground
truth data. To train the random trees image classifier, 500 random samples of individual pixels
were selected through classification results and were manually compared for accuracy against the
ground truth data, which is the 30 cm resolution satellite imagery. The matrix displayed in Figure
12 also calculates the Kappa Statistic Index, which is the score for the overall accuracy of the

47
supervised image classification based on the Type 1 and Type 2 errors that were identified. For
the random trees algorithm, the overall accuracy score was 0.553408 or 55.3408% accuracy,
which means that this algorithm only classified just over half of the pixels correctly.
4.1.2 Support Vector Machine
The other algorithm that is employed during the supervised image classification process
was the support vector machine algorithm. Additionally, the input for the SVM supervised image
classification is the 30 cm aerial imagery acquired from Land Info, as displayed in Figure 13.
The imagery found in Figure 13depicts the Pueblo Bonito ruin, which is the most famous of all
the ruins at Chaco Canyon.
Figure 12. Random Trees Algorithm Confusion Matrix

48

Figure 13. Satellite imagery at 30 cm resolution from Pleiades Satellite
To provide a frame of reference, the high-resolution imagery of Pueblo Bonito found in
Figure 13 is in the same position as the SVM results displayed in Figure 14. The same color
classification scheme is used as the previous supervised image classification employing the
random trees algorithm. The barren desert is classified as yellow, vegetation as green, ruins as
red, and ancient pathways as orange, purple for developed land, and gray for asphalt (which
often gets confused with shadows in the imagery, as displayed later in Figure 19). Comparing the
RT results, the SVM algorithm did a much better job as it correctly classified more pixels. The
structural footprint of Pueblo Bonito is clearly visible, and the results were determined to be
more accurate than the RT results according to the confusion matrices for the two classifications.
The SVM image classification results are more useful to this research when determining the
shape, size, and location of the archaeological ruins and ancient roadways in Chaco Canyon.

49
These results also help assess the physical condition of site Bc 53 based on whether it is in a
good enough state of preservation to be identified during the supervised image classification.
Figure 14. Classification of Pueblo Bonito employing SVM Algorithm
To evaluate the accuracy of the SVM-supervised image classification results, a confusion
matrix is generated that compares the classification results against the ground truth data, as
shown in Figure 15. To train the SVM image classifier, 500 individual pixels are selected as
samples, and each pixel's classification result is compared to the ground truth data, which for this
project is satellite imagery with a spatial resolution of 30 cm. The Kappa Statistic, or the overall
accuracy for the SVM supervised image classification, is 0.814825 or 81.4825%, which is
significantly better than the overall accuracy of the random trees algorithm, which has an overall
accuracy of 55.3408%. While 81.4825% is not a perfect accuracy score, the SVM algorithm did
classify over 80% of the pixels correctly, making a product that can be utilized to conduct an

50
exploratory analysis of the Chaco Canyon study area searching for any new anomalies, whether
ruins or ancient roadways, that are not distinguishable with the naked eye. Because the SVM
image classification results have an overall accuracy of over 80%, the majority of the known
ruins are correctly classified, as displayed in Figure 14. In addition to being used for the
exploratory analysis, the SVM classification results are used to assess the preservation state of
site Bc 53. The better preserved the ruin, the more distinguishable the outline of the ruin in the
supervised image classification results.
Figure 15. Support Vector Machine Algorithm Confusion Matrix
4.2 Significance of Site Bc 53
In this section, the primary focus is on Site Bc 53 and the LCPs that are generated from
Site Bc 53 to its neighboring sites. First, the results of the LCP analysis between Site Bc 53 and
the nearby Casa Rinconada community are explored with the aim of revealing the connection
between the two sites. Subsequently, the results of the LCP analysis between Site Bc 53 and the
FP Chacoan Staircase are examined, and the relationship between the two sites is further
investigated.

51
4.2.1 LCP Analysis Between Site Bc 53 and the Nearby Casa Rinconada Structures
As illustrated in Figure 16, the input raster for the LCP analysis is the LiDAR-derived
DEM, which was obtained from OpenTopography at a spatial resolution of 1-meter. The teal
cross, which is also visible in Figure 17, represents the location of site Bc 53, while the red circle
denotes the position of the Casa Rinconada ruins. These two points serve as the source and
destination input points when conducting the LCP analysis in ArcGIS Pro. The primary objective
of this analysis is twofold: to determine the feasibility of the paths between site Bc 53 and the
Casa Rinconada community of structures and to uncover whether site Bc 53 should be included
in the Casa Rinconada community of ruins group or if Bc 53 should be categorized as
independent.
Figure 16. Points for the LCP Analysis and LiDAR-derived DEM input raster

52
Figure 17 displays the outcome of the LCP analysis, illustrating the path deemed most
optimal based on terrain and vegetation conditions, connecting one of the structures within the
Casa Rinconada community (site Bc 51) to the nearby site Bc 53. The source point for the LCP
analysis is site Bc 53, which is marked by the teal-colored cross. The destination of this LCP
analysis is one of the Casa Rinconada ruins, specifically the ruin closest to site Bc 53. The green
line identifies the most optimal path between the two locations, and it is based on the terrain at
Chaco Canyon. For this LCP analysis, terrain is the main variable driving the LCP analysis
because terrain is the primary factor that influenced Anasazi trade routes, short-distance travel
networks, and long-distance travel networks.
In the 9th through the 12th centuries CE, the Anasai were limited to traveling by foot,
meaning that their pathways, both for long distances and local travels, were heavily impacted by
the natural landscape. The Anasazi constructed their road network to align with the natural
Figure 17. Least cost path between Bc 53 and Casa Rinconada Community

53
contours of the landscape, ensuring efficient travel without expending excessive energy on a
regular basis. Therefore, terrain was chosen as the main factor for the LCP analysis. Analyzing
topography and terrain is a crucial initial step in understanding the movement patterns of the
Anasazi people.
However, it's crucial to note that human behavior may not always align with
environmental factors alone. Ancient travel patterns could have been influenced by religious or
social customs that are not able to be captured by GIS software. Therefore, it's essential to
consider a range of external variables, drawing from prior archaeological research on Anasazi
society and cultural practices, to further enhance the LCP analysis in future research. Some other
variables that were considered for this research were creating barriers around fertile land that
were most likely used for agriculture and not travel. Another variable that was considered for this
analysis was assigning weights to areas that have been documented to be trade corridors for the
Anasazi. It was also considered to customize the LCP analysis to the individuals who were
traveling and how their demographics and behavior influenced their travel routes.
The results depicted in Figure 17 suggest a viable route between site Bc 53 and the Casa
Rinconada community of ruins. However, due to the indirect nature of the most optimal path, it's
plausible that Bc 53 served as a neighboring site rather than being directly integrated into the
community itself. Its relatively isolated location beneath the butte further supports this
interpretation.
4.2.2 LCP Analysis Between Site Bc 53 and the Nearby FP Chacoan Staircase
Another LCP analysis is conducted between site Bc 53 and the nearby FP Chacoan
Staircase, as depicted in Figure 18. Site Bc 53 is marked with the green cross and serves as the
source point when running the LCP analysis. A red octagon marks the base of the FP Chacoan

54
Staircase and serves as the destination point when completing the LCP analysis. The yellow line
between the source and destination points is the most optimal path based on the terrain at Chaco
Canyon that was extracted from the 1-meter LiDAR-derived DEM. These results suggest that a
direct relationship between Bc 53 and the FP Chacoan Staircase is feasible based on the
proximity and accessibility between the two sites; however, in order to definitively link these
two sites together, more in-depth research that considers more variables than the terrain is
necessary.
The LCP analysis provides valuable insights into the movement patterns and daily travel
routes within Chaco Canyon, particularly between various sites and their connections to the
ancient roadway network. The LCP analysis reveals the most optimal paths for short, routine
travel, offering a complementary perspective on Anasazi mobility distinct from the formalized
short- and long-distance roadways. The potential routes identified through the LCP analysis
Figure 18. Least cost path between Bc 53 and the FP Chacoan Staircase

55
represent the practical routes that individuals may have taken, shaped by the terrain and the need
for efficiency in daily activities.
These footpaths, however, differ significantly from the engineered short- and longdistance roadways, which were substantial, labor-intensive constructions designed for durability
as well as ceremonial and symbolic purposes. The Chacoan roadways, known for their width and
straightness, required considerable resources to build and maintain, reflecting their importance in
connecting distant communities and facilitating large-scale movement across the region. In
contrast, the potential paths revealed by the LCP analysis may have been created organically
through repeated use. They were not intended to be permanent features, which may explain why
they are not as visible today as the more formalized road networks.
4.3 Assessment of the Physical Condition of Site Bc 53
This section focuses on using the LiDAR data of Chaco Canyon, as well as the results
from the supervised image classification to investigate the preservation state of site Bc 53 in
comparison to the other nearby ruins. The section first explores the results of the SVM
supervised image classification, specifically examining the location of Site Bc 53 with the aim of
seeing if any of the ruins are indefinitely in the supervised image classification. The section
concludes by using the LiDAR data acquired for this project to see if any part of site Bc 53 is
identifiable, just as the other nearby ruins are.
4.3.1 Assessing Site Bc 53 Using Supervised Image Classification Results
As part of this research, the physical condition of site Bc 53 is investigated utilizing
remote sensing technology and advanced GIS techniques. During the supervised image
classification of Chaco Canyon in its entirety, almost all of the ruins were successfully identified
based on their spectral signatures in the high-resolution imagery, as shown in Figure 11 and

56
Figure 14. The Casa Rinconada community of structures and site Bc 53 were occupied around
the end of the 11th century and therefore should have used the same building materials. The use
of sandstone bricks, mud mortar, and adobe in construction means that the spectral signature of
site Bc 53 should be identifiable just like the Casa Rinconada community due to the shared
building materials. However, when examining site Bc 53, no definitive outline of the ruin could
be identified, as shown in Figure 19. The blue circle denotes the area where site Bc 53 is located.
These results suggest that the physical condition of site Bc 53 is rougher than that of the other
ruins in the area, which were easily identifiable in contrast to the barren desert landscape
Figure 19. Supervised image classification (SVM) of Bc 53
4.3.2 Assessing Site Bc 53 Using LiDAR-Derived DEM
To validate the findings of the supervised image classification, the LiDAR-derived
Digital Elevation Model is also inspected to determine site Bc 53's physical condition when
compared to the other ruins at Chaco Canyon. The inspection was carried out at a spatial

57
resolution of 1-meter based on the spatial resolution of the LiDAR data acquired from
OpenTopography. As shown in Figure 20, the blue circle represents the area where there should
be some evidence of the ruin, but no structural footprint is detectable. In contrast, the nearby
Casa Rinconada community of ruins which are identified by the red circles. These results
confirm that site Bc 53 is not in good physical condition, as even when examining the LiDARderived DEM, it was still not clearly identifiable compared to the other nearby ruins.
As depicted in Figure 21, the ruins that make up the Casa Rinconada community of
structures are easily identifiable in contrast to the barren desert landscape. In Figure 21, the ruins
that make up the Casa Rinconada community are identified by the red circles. The blue arrow
denotes the main structure of the Casa Rinconada community and is also the most defined ruin in
the DEM. This makes sense because it is the most well-preserved of all the Casa Rinconada
Figure 20. LiDAR-derived DEM of Bc 53 and Casa Rinconada Community

58
ruins. Further following this logic, the level of detail and distinguishability of these nearby
structures in the DEM can be used as the baseline to evaluate the preservation condition of the
nearby site Bc 53. If site Bc 53 is as well preserved as the ruins in the Casa Rinconada
community, then the outline of site Bc 53 should have similar discernibility in the DEM.
Figure 22 displays a zoomed-in view of the LiDAR-derived DEM focused on site Bc 53
and the edge of the Casa Rinconada community of ruins. As shown in Figure 22, the Casa
Rinconada ruins are clearly identifiable and are circled in red. In contrast, site Bc 53 is circled in
blue, and the outline of the ruin is not identifiable. In fact, the location of site Bc 53 appears to be
one large pit. This could be the result of the initial and destructive excavation that was then
backfilled with the excavated sediment after the excavation was completed. This DEM further
Figure 21. LiDAR-derived DEM of the Casa Rinconada Community

59
confirms that site Bc 53 is in poor preservation condition in comparison to the nearby ruins.
These results also highlight the need for further exploration into site Bc 53, with the aim of
uncovering why the footprint of the ruin is completely unrecognizable in comparison to the other
nearby sites and if there are any preservation techniques that can be employed to combat the
complete deterioration of site Bc 53.
4.4 3D Visualization of the LiDAR-Derived Digital Elevation Model
This final section of the results chapter analyzes the process of generating the final
contiguous 3D visualization of Chaco Canyon. It first discusses how the LiDAR-derived DEM is
transformed into a usable product for this research. This section then details how the
visualization is made realistic by incorporating high-resolution satellite imagery. The symbology
Figure 22. LiDAR-derived DEM of Bc 53

60
of the ruins, ancient roadways, and the LCP analysis results are reviewed. This section concludes
by outlining how the 3D visualization was further transformed into a 3D Web Scene on ArcGIS
Online for public viewing.
4.4.1 Chaco Canyon Visualized in a 3D Scene in ArcGIS Pro
As displayed in Error! Reference source not found., the 3D scene of Chaco Canyon
was created in ArcGIS Pro using the Digital Elevation Model derived from the LiDAR point
cloud acquired from OpenTopography. In the forefront of Error! Reference source not found.,
the Chaco Wash and Pueblo Bonito ruin can both be seen. The previously collected LiDAR data
did an excellent job of capturing the nuances in the topography and highlighting the different
paths and roadways that may not be visible to the naked eye. This DEM is the input for
generating the 3D visualization of Chaco Canyon.

61
4.4.2 Combining the 3D Visualization with High-Resolution Satellite Imagery
Figure 23 is a screen capture of the 3D visualization generated in this project facing
Pueblo Bonito and Pueblo del Arroyo. A golden star symbol designates the major ruins in the
visualization of Chaco Canyon. Figure 24 is also a screen capture of the 3D visualization
generated for this research; however, it faces the Casa Rinconada community of ruins and site Bc
53. The orange pathways that are illustrated in both Figure 23 and Figure 24 are the ends of the
short-distance roadway network that ends at Downtown Chaco Valley bottom.
Figure 23. Generated 3D visualization of Chaco Canyon facing Pueblo Bonito

62
In Figure 25, the results of the LCP analysis are illustrated along with the source and
destination points. The yellow path is the most optimal path between the source, site Bc 53,
marked with a green cross, and the destination point, which is the base of the FP Chacoan
staircase, marked by a red octagon. The path in purple is the most optimal path between the
source, site Bc 53, marked with a green cross, and the destination point, which is the edge of the
Casa Rinconada community of structures (site Bc 51) and marked by a red octagon. The orange
paths are the edges of the short-distance ancient roadway network, which will be discussed
further in the research.
Figure 24. Generated 3D visualization of Bc 53 and Casa Rinconada Community

63
The 3D visualization is first created by creating the multidirectional hillshade layer from
the LiDAR-derived DEM. The 2D map is then transformed into a local scene, which switches
the map and all of its layers to a 3D view. The LiDAR-derived DEM layer is then dragged under
elevation surfaces to replace the default global terrain. Once the visualization is in 3D, the
satellite imagery that is wrapping the visualization can now be adjusted to achieve maximum
realism. For this project, the Layer Blending and Overlay tools were utilized in ArcGIS Pro to
combine the 3D visualization of Chaco Canyon with the high-resolution satellite imagery (also at
a spatial resolution of 1-meter). Figure 23 shows the results of the generated 3D visualization
facing towards the Pueblo Bonito and Pueblo de Arroyo ruins, while Figure 24 shows the results
of the 3D visualization facing towards the Casa Rinconada community and site Bc 53. The
results of the LCP analysis are added to the 3D scene, as displayed in Figure 25, and the major
ruins are identified with the golden star symbol by creating individual feature points. In order to
Figure 25. LCP analysis results displayed in the 3D visualization

64
visualize the ancient roadway network in ArcGIS Pro, the map created by Lekson in 1988 is used
as a reference to illustrate the short-distance roadway network, as displayed in Figure 27, and the
long-distance roadway network, as displayed in Figure 28. The map was first brought into
ArcGIS Pro as a .jpg file and then georeferenced using common “tie points” that were
distinguishable in both the reference map and the satellite imagery for this project. Line features
were then created and sketched over the georeferenced Lekson reference map to generate both
the short-distance and the long-distance roadway networks as identified in Lekson's research
(Lekson 1988).
Figure 26 illustrates the final 3D visualization generated for this research project. It is a
zoomed-out view of the Casa Rinconada community of ruins and site Bc 53. In this figure, the
short-distance roadway network is depicted as orange-colored trails. Towards the center of
Figure 26, the LCP analysis results are depicted along with the major archaeological sites
designated by golden stars. The two destination points for the LCP analysis are red octagons, site
Bc 51, and the FP Chacoan staircase, and the source point, site Bc 53, is marked by a green
cross.

65
Figure 27 displays an aerial view of the final 3D visualization generated for this research
project, focusing on the short-distance roadway network. The short-distance ancient roadway
network created by the Anasazi is illustrated in orange, which connects with the long-distance
ancient roadway network, which is illustrated in the color teal. Toward the center of Figure 27,
golden stars designate the locations of the major ruins in Chaco Canyon National Historical Park.
The final features that are depicted in Figure 27 are the elements of the LCP analysis, site Bc 53,
which is a green cross, and the two destination points as red octagons, site Bc 51, and the FP
Chacoan staircase.
Figure 26. Final 3D visualization of Chaco Canyon zoomed-out

66
Figure 28 depicts the Four Corners region in the American Southwest and highlights the
extensive ancient long-distance roadway network created by the Anasazi. The long-distance road
network is shown in teal towards the center of Figure 28. An orange-colored mass represents the
short-distance roadway network displayed in Figure 28, illustrating its connection to the longdistance network. In the middle of the orange mass, a red octagon denotes one of the destination
points from the previous LCP analysis. Another important feature in Figure 28 is the faded pink
roadway network, which actually represents the modern-day highway system in the American
Southwest. Remarkably, this modern highway system closely resembles the ancient longdistance roadway network in teal. This similarity is a testament to the impressive ingenuity of the
Anasazi and their advanced thinking for their time. The final contiguous 3D visualization of
Figure 27. Short distance ancient roadway network

67
Chaco Canyon is available on ArcGIS Online as a 3D Web Scene at the following link:
https://www.arcgis.com/home/item.html?id=9d984e5f734f44ef83d9b5fa72e5b88b
4.4.3 Ethical Implications of the Final 3D Visualization
The decision to make the 3D visualization of Chaco Canyon publicly accessible on
ArcGIS Pro raises important ethical considerations, particularly regarding the protection of
cultural heritage and the potential for looting. While the visualization has significant educational
Figure 28. Long distance ancient roadway network

68
and research value, it is essential to address these concerns within the broader context of making
archaeological site locations publicly available.
One important factor taken into consideration when deciding to make this 3D
visualization publicly accessible is the abundance of information about the sites at Chaco
Canyon that is already available on widely used platforms such as Google Earth, Google Maps,
and the Hybrid Satellite Imagery basemap in ArcGIS Pro. These platforms already provide
detailed locations for most of the ruins within Chaco Canyon. These platforms deliberately blur
the locations of sacred sites, such as Tsin Kletsin, to protect them from being easily located. The
visualization created for this project does not introduce additional risks beyond what is already
publicly accessible. This research also ensures that the sacred sites are blurred for their continued
protection and privacy.
To further mitigate any potential risks, the National Park Service archaeological team at
Chaco Canyon was consulted during the development of this project. Their support highlights the
value of the visualization since the National Parks Service does not currently possess a complete,
contiguous 3D visualization of the area. This collaboration ensured that the project aligned with
the NPS’s preservation goals. Additionally, the project proposal was submitted to the Tribal
Review Board by the NPS team, which carefully considered its cultural implications. Although
certain proposed research activities, such as using a handheld LiDAR scanner to create a 3D
model of site Bc 53, required additional discussion and approval, the objective of generating the
3D visualization was deemed acceptable.
Chapter 5 Conclusions and Discussion
In this conclusion chapter, there are two main areas of focus. The first section breaks down the
results of the research project and elaborates on their significance. The final section provides an

69
objective overview of the research, identifying major limitations and areas for improvement.
This chapter concludes by discussing directions for future research.
5.1 Deciphering the Results
This section provides a detailed analysis of the outputs from the three primary
methodologies employed for this research. The first subsection addresses the results of the
supervised image classification, highlighting its initial objectives, the achieved accuracy, and the
implications of these results for identifying archaeological sites. The following subsection
focuses on the analysis of Site Bc 53 and uses the LCP analysis to predict the movement and
connectivity patterns of the Anasazi between this site and its neighbors. The final subsection
discusses the 3D visualization product, explaining how the generated 3D model integrates terrain
data, satellite imagery, and archaeological site markers to create a comprehensive and navigable
representation of Chaco Canyon and its ancient roadway networks. The goal of this section is to
provide a comprehensive understanding of the research findings and their implications for future
studies in archaeological site detection and analysis.
5.1.1 Supervised Image Classification Results
The results obtained from the supervised image classification did not meet the original
expectations established at the beginning of this research. The initial goal of the supervised
image classification was to create a machine-learning algorithm that could accurately identify
concealed ruins with the same building material as the known ruins. If the same building
materials were used to construct the hidden ruins as the known ruins, then theoretically, they
would have the same spectral signature. Unfortunately, even when trying two separate
algorithms, the highest overall accuracy score achieved was just over 80%, meaning that the
image classification only classified 80% of the pixels correctly. While the supervised image

70
classification algorithm was not accurate enough to detect unknown ruins definitively, it was still
able to detect the known ruins around Chaco Canyon National Historical Park. The objective of
the supervised image classification was then adjusted to utilize the results of the image
classification to investigate why site Bc 53 was not distinguishable in the image classification
like the other nearby ruins.
The results of the supervised image classification were valuable in validating the
methodology for using a supervised image classification to identify archaeological sites and
ancient roadways. While this initial application of supervised image classification was not
perfect, it was able to detect known archaeological sites. This initial research provides an
opportunity to enhance the methodology of this image classification process further by
increasing the number of training samples and conducting the image classification on satellite
imagery with even higher spatial resolution in the future. By enhancing the accuracy and
robustness of the supervised image classification, the methodology developed for this research
has great potential to identify previously unknown archaeological areas of interest successfully.
This classification provides a solid foundation for future research to conduct more
comprehensive supervised image classification, particularly as the spatial resolution of satellites
and other aerial technology continues to improve, leading to increasingly accurate results with
each enhancement in spatial resolution.
5.1.2 Site Bc 53 Analysis Breakdown
The main goal of the LCP analysis was to determine the most plausible pathways taken
by the Anasazi between site Bc 53 and its neighboring sites. The LCP analysis results
demonstrate the most efficient routes between the project's source point, Site Bc 53, and the
neighboring FP Chacoan Staircase and Site Bc 51. These results illustrate the directness of the

71
optimal path between Site Bc 53 and the nearby sites, revealing the relationship between the sites
based on their proximity and accessibility to one another. This LCP analysis is critical for
understanding how the Anasazi people navigated the canyon landscape because it uses empirical
evidence about the landscape, such as slope, vegetation cover, and natural barriers, to determine
the most plausible local pathways created by the Anasazi and visualize them.
By mapping the LCPs between Site Bc 53 and the nearby destination points, this project
identified the most probable routes that the Anasazi would have taken between individual sites to
minimize effort during travel. This LCP analysis illuminates the daily movement of the Anasazi
within the canyon and enhances the understanding of the strategic connections between these
sites. Understanding the results of the LCP analysis offers valuable insights into the cultural and
functional significance of various sites, including the utilization of natural features in their
architecture, such as the FP Chacoan Staircase, to connect different roadway networks.
This analysis offers data on potential pathways linking Bc 53 with nearby archaeological
sites, providing insight into their relationships, such as their spatial proximity and how the sites
are physically connected. Additionally, there is limited research on the local pathways within the
canyon that connect less prominent sites like Bc 53. Although Bc 53 shares the same “Bc” name
as the nearby ruins of the Casa Rinconada community, the National Historical Park determined
Bc 53 to be separate from the Casa Rinconada community of structures. The findings of the LCP
analysis conducted between site Bc 53, and the Casa Rinconada ruins (site Bc 51) suggest that
the most optimal route between the two sites is not direct. The indirect nature of the most optimal
path implies that Bc 53 may have been a neighboring site rather than fully integrated into the
community. The relatively isolated position of site Bc 53 beneath the butte supports this
interpretation.

72
The results of the LCP analysis between site Bc 53 and the FP Chacoan staircase indicate
that there could be a possible connection between them due to their close proximity and the more
direct nature of the most optimal path between them. However, further research that is more
robust and considers additional variables beyond just the terrain is required to conclusively
establish a link between these two sites.
While the LCP analysis for this research is only driven by the topography that was
extracted from the 1-meter DEM, it still provides a solid foundational analysis that can be
enhanced to incorporate more variables when determining the most optimal paths. For this
research, some other variables that were considered were designating areas in the analysis as
fertile land more likely used for agriculture rather than travel. Another consideration was
assigning weights to the known trade routes of the Anasazi. The final consideration was to focus
the analysis on the demographics and behavior of the Anasazi travelers and how these factors
influenced their travel routes. An Agent-Based Model may be better equipped than an LCP
analysis to predict ancient pathways based on individual human behaviors and cultural customs.
An Agent-Based Model is an alternative advanced GIS analysis that models dynamic
geographical phenomena using individual “agents.”
When evaluating the preservation condition of site Bc 53 during this research using the
LiDAR-derived DEM, it becomes evident that Site Bc 53 lacks any visible outline of the
structure in contrast to some of the nearby ruins, such as Bc 50 and Bc 51. This anomaly was
discussed with Lori Stephens, the Supervisory Archaeologist at Chaco Canyon, who explained
that Chaco Canyon National Historical Park has a team of preservation masons who conduct
stabilization work on sites with standing architecture. Site Bc 53 is on the NPS list for cyclic
preservation work. However, Lor Stephens explained that there are only a few sections of wall

73
exposed on the west side of the site that are taller than 3 ft high. The limited exposure of Bc 53 is
confirmed by the lack of visibility of the structure during the supervised image classification and
the examination of the LiDAR-derived DEM for this research. Lori Stephens also addressed the
previous question about the absence of public trails allowing access to site Bc 53. She explained
that the park management, when establishing the trail system, made decisions about trail
placement and accessibility for the public. This balance between preserving and protecting
resources and educating the public on these amazing sites necessitates careful consideration of
which sites to protect by limiting access and where to create trails through sites that would be
open to the public. Specifically, around Site Bc 53, there are no trails at all leading to or around
the site, making it essentially closed off to the public.
It is initially alarming that there appears to be a pit of loose soil in the area where Site Bc
53 was determined to be located. However, Supervisory Archaeologist Lori Stephens provided
assurance that Site Bc 53 is, in fact, still protected by soil that was backfilled after the initial
excavation in the 1940s. Nonetheless, it is exciting that the LiDAR-derived DEM was able to
identify this anomaly in the midst of the otherwise uniform desert landscape. While some ruins at
Chaco Canyon are left exposed to the elements, such as Pueblo Bonito and the Casa Rinconada
structures, the best plan of action at the time to protect Site Bc 53 was to backfill the site. Lori
explained that when some sites are excavated and the walls are exposed to the elements,
deterioration of the architecture is a significant concern. As a result, many of the excavated sites
in the park were backfilled with soil after excavation to protect the remaining structure, and Site
Bc 53 was one of these sites.
Additionally, any sites that have standing architecture along visitor trails require a
significant effort by their preservation masonry crew to keep the stone and mortar stable and

74
prevent deterioration. The National Historical Park decided that some sites would remain closed
to the public, and the exposed architecture would be backfilled to protect them and prevent
deterioration of the walls. It still remains unclear why Site Bc 53 was backfilled in contrast to the
other nearby “Bc” sites that make up the Casa Rinconada community of structures such as Bc 50
and Bc 51. Unfortunately, now that the ruin is backfilled, it is not possible to evaluate the quality
of the archaeological work conducted in the 1940s' without conducting a complete re-excavation.
For now, preserving the structure in its current state is the best course of action.
5.1.3 3D Visualization Product and Results
The finalized 3D model of Chaco Canyon is now available to the public on ArcGIS
Online as a 3D Web Scene. The visualization was initially created in ArcGIS Pro, then published
as a 3D Web Scene, and further edited in ArcGIS Online. The terrain for the 3D model is based
on a LiDAR-derived DEM with a spatial resolution of one meter. The 1-meter satellite imagery
was used to wrap the generated 3D model, providing a seamless match with the terrain for the
3D visualization. In the visualization, golden star symbols indicate the major ruins, and users can
easily navigate from one archaeological site to another using pre-determined “slides” of each
point of interest created in the Slide Manager in ArcGIS Online. A green cross identifies Site Bc
53, and the two destination points for the LCP analysis are site Bc 51 and the FP Chacoan
Staircase, which are marked by red octagons. Both of the optimal paths determined during the
LCP analysis are also illustrated in the final 3D visualization, as displayed in Figure 25. The
ancient short-distance roadway network is depicted as orange pathways, aiding in understanding
the purpose of these roadways and how they connect with the local pathways within the canyon,
as well as with the long-distance roadway network. The ancient long-distance roadway network
is illustrated in teal in the visualization and depicts the expansive nature of the Chacoan roadway

75
network spanning across the Four Corners Region of the American Southwest. As displayed by
the connection of the orange and teal roadways, it is evident that the Anasazi had a hierarchy of
roadways, ranging from local pathways to short-distance travel and extending to long-distance
travel.
5.2 Evaluating the Overall Research and Future Directions
This section discusses the main findings and limitations encountered during the
supervised image classification, LCP analysis, and 3D visualization processes applied in this
thesis research project on the Chaco Canyon archaeological sites. Specific challenges related to
the accuracy of the image classification, the limitations of the LCP analysis in capturing
culturally specific factors, and the issues of distortion in 3D visualization are addressed.
Additionally, potential directions for future research that could improve the accuracy and
practicality of these methods are outlined, aiming to contribute to a more comprehensive
understanding of the ancient wonders at Chaco Canyon.
5.2.1 Research Limitations
The main limitation of the supervised image classification process during this research
was the relatively low accuracy score. While the 80% accuracy of the SVM classification was
still useful for this project, it would have been ideal to have achieved a higher overall accuracy
score to assist with the intended purpose of this project of definitively distinguishing
undocumented ruins amongst the desert landscape. Future research that builds upon this study
should work with imagery that offers a higher contrast, which will be better equipped to pick up
on all the different spectral signatures in the imagery and to easily distinguish between categories
with similar spectral signatures, such as barren desert and ancient roadways. Additionally, if
more training data is collected for each category in the ArcGIS Pro training sample manager, it

76
would allow for the image classification to do a better job with class distinction, which
ultimately leads to better overall detection accuracy.
The LCP analysis for this thesis research only considered the terrain and did not
incorporate other variables, so the results are not as tailored for the Anasazi as originally
intended. Due to the limited number of variables in the LCP analysis, the results were not as
meaningful as intended, and they failed to consider cultural, social, or religious factors when
determining the most likely pathways taken by the Anasazi.
The primary limitation of the 3D visualization is that when the satellite imagery is
applied to the generated 3D model, some areas of the imagery become distorted, particularly in
areas with steep cliffs or drop-offs. This distortion of the imagery on steep terrain is not a
significant problem for most of the major archaeological sites. However, for sites like the FP
Chacoan Staircase, which is built into the cliffside, the distortion almost completely obscures the
imagery.
5.2.2 Future Research
Even though the SVM image classification was able to correctly identify over 80% of the
pixels, future research could improve its accuracy by running the SVM image classification
using more training samples and different satellite imagery with even higher spatial resolution.
This increased accuracy could help create a machine-learning model that can assist
archaeologists in identifying ancient structures and roadways beyond the current boundaries of
Chaco Canyon quickly and effectively.
While there are limitations to the LCP analysis completed for this research, recognizing
these limitations, and building upon this foundational layer presents an opportunity for future
research to incorporate more culturally specific variables or environmental factors. This could

77
lead to a more accurate understanding of the ancient roadway system in Chaco Canyon. Starting
with a topography-based LCP analysis establishes a baseline understanding of movement in
Chaco Canyon that can be further refined for more in-depth results in future research.
By conducting a brief analysis of Site Bc 53, the aim is to pave the way for future
archaeological investigations in similar contexts. These investigations can utilize remote sensing
technology and advanced GIS techniques to re-evaluate previously excavated sites for additional
insights that may have been overlooked initially. Remote sensing is a valuable tool for
archaeologists, providing a wide range of applications to extract data about an archaeological site
without physical interaction. Due to the non-invasive nature of remote sensing technology,
archaeologists have the opportunity to create digital models of archaeological sites, allowing for
efficient collaboration, repeated investigations, and the production of 3D visualizations that can
highlight and contextualize archaeological findings a change from prior archaeological
excavation methodologies that were more invasive and less sustainable.
The 3D visualization developed for this project can be used for future research,
conservation efforts, educational purposes, and to inform policy decisions regarding the
management of significant national and international heritage sites like Chaco Canyon. This
visualization serves as a solid basis for future research that can delve deeper into the extensive
ancient roadway network or map the locations of ancient structures outside the park boundaries
that are under threat from modern risks such as encroaching oil and gas drilling. Additionally,
the goal of this 3D visualization is to inspire other archaeologists and protection agencies, such
as the National Parks Service, to allocate time and resources to create 3D models to “digitally
protect” their archaeological sites for future research and exploration.

78
References
Adomaitis, C. “The Navajo Nation Comes Out Against the Interior Department's Buffer Zone
Around Chaco Canyon.” NPR Four Corners Public Radio., last modified June 9, accessed
03/11/, 2024, https://www.ksut.org/news/2023-06-09/the-navajo-nation-comes-outagainst-the-interior-departments-buffer-zone-around-chaco-canyon.
Alvez, C. 2016. “Modeling Prehistoric Paths in Bronze Age Northeast England.” Master’s
Thesis, Spatial Sciences Institute, University of Southern California.
Argyrou, A., A. Agapiou, A. Papakonstantinou, and D. Alexakis. 2023. “Comparison of Machine
Learning Pixel-Based Classifiers for Detecting Archaeological Ceramics.” Drones 7: NA.
https://link.gale.com/apps/doc/A771804924/AONE?u=anon~9651fd9a&sid=sitemap&xi
d=773588d7.
Basso, K. H. 1996. “Wisdom Sits in Places Landscape and Language among the Western
Apache.” University of New Mexico Press. https://nimshav.github.io/EthnoCommRepository/EOC_Library/Basso%20-%201996%20-
%20Wisdom%20sits%20in%20places.pdf.
Budge, T., A. Komarek, and S. Morain. 1981. Remote Sensing: Multispectral Analyses of
Cultural Resources, Chaco Canyon and Bandelier National Monument, edited by Thomas
Lyons. Library of Congress ed. Washington DC: Cultural Resource Management
National Park Service US Department of the Interior.
Carter, W., J. Carlos Fernandez-Diaz, R. Shrestha, and C. Glennie. 2014. “Now You See It…
Now You Don’t: Understanding Airborne Mapping LiDAR Collection and Data Product
Generation for Archaeological Research in Mesoamerica” Remote Sensing 6, no. 10:
9951-10001. https://doi.org/10.3390/rs6109951
Casana, J. 2020. “Global-Scale Archaeological Prospection using CORONA Satellite Imagery:
Automated, Crowd-Sourced, and Expert-Led Approaches.” Journal of Field Archaeology
45: S89-S100. doi:10.1080/00934690.2020.1713285.
https://doi.org/10.1080/00934690.2020.1713285.
Chaco Research Archive, Institute for Advanced Technology in the Humanities and the
Department of Anthropology, University of Virginia. “Bc 53 – Roberts’ Site.” Bc 53 –
Roberts’ Site., accessed 02/09/, 2024, http://www.chacoarchive.org/cra/chaco-sites/bc53-roberts-site/.
Corns, A. and Robert S. 2009. “High Resolution 3-Dimensional Documentation of
Archaeological Monuments & Landscapes using Airborne LiDAR.” Journal of Cultural
Heritage 10: e72-e77. doi:10.1016/j.culher.2009.09.003.
https://www.sciencedirect.com/science/article/pii/S129620740900137X.

79
Ditto, E. 2017. “Cosmological Caches: Organization and Power at Chaco Canyon, New Mexico,
A.D. 850-1150.”The University of North Carolina at Chapel Hill.
https://www.proquest.com/docview/1952049146?pqorigsite=gscholar&fromopenview=true&sourcetype=Dissertations%20&%20Theses#
Dorshow W. 2010. “Chaco Canyon, NM: Simulating Dynamic Hydrological Processes.”
OpenTopography., accessed 02/09/, 2024,
https://portal.opentopography.org/lidarDataset?opentopoID=OTLAS.112011.26913.1.
Field, S. 2023. “Lidar-Derived Road Profiles: A Case Study Using Chaco Roads from the US
Southwest.” Advances in Archaeological Practice 11, no. 2 (2023): 184–97.
https://doi.org/10.1017/aap.2022.31
Field, S., C. Heitman, and H. Richards-Rissetto, 2019. “A Least Cost Analysis: Correlative
Modeling of the Chaco Regional Road System. Journal of Computer Applications in
Archaeology,” 2(1), pp.136–150. DOI: http://doi.org/10.5334/jcaa.36
Friedman, R., A. Sofaer, and R. Weiner. 2017. “Remote Sensing of Chaco Roads Revisited:
Lidar Documentation of the Great North Road, Pueblo Alto Landscape, and Aztec
Airport Mesa Road.” Advances in Archaeological Practice 5, no. 4 (2017): 365–81.
https://doi.org/10.1017/aap.2017.25.
Friedman, R., A. Sofaer, and R. Weiner. 2021. “LiDAR and 3-D Digital Modeling Reveal the
Greater Chaco Landscape.” In The Greater Chaco Landscape, edited by Ruth M. Van
Dyke and Carrie C. Heitman, 229-255: University Press of Colorado.
http://www.jstor.org/stable/j.ctv1m46ffr.16.
Gellis, A. 2002. “Twentieth Century Arroyo Changes in Chaco Culture National Historical
Park.” United States Geological Survey. doi:https://doi.org/10.3133/wri014251.
https://pubs.usgs.gov/publication/wri014251.
Gumerman, G. and J. Ware. 1977. “Remote Sensing Methodology and the Chaco Canyon
Prehistoric Road System.” In Aerial Remote Sensing Techniques in Archeology, edited
by Thomas R. Lyons and Robert K. Hitchcock, 135-167. University of New Mexico
Albuquerque: Chaco Center National Parks Service.
Guo, M., Z. Fu, D. Pan, Y. Zhou, M. Huang, and K. Guo. 2022. “3D Digital Protection and
Representation of Burial Ruins Based on LiDAR and UAV Survey.” Measurement and
Control 55 (7-8): 555-566. doi:10.1177/00202940221110949.
https://doi.org/10.1177/00202940221110949.
Judge, W., S. Lekson, T. Windes, J.. Stein, 1988. “The Chaco Canyon Community.” Scientific
American 259 (1): 100-109. http://www.jstor.org/stable/24989164.
Katsianis, M., S. Tsipidis, K. Kotsakis, and A. Kousoulakou. 2008. “A 3D Digital Workflow for
Archaeological Intra-Site Research using GIS.” Journal of Archaeological Science 35 (3):
655-667. doi:10.1016/j.jas.2007.06.002.
https://www.sciencedirect.com/science/article/pii/S0305440307001069.

80
Keeney, J., and R. Hickey. 2015. “Using Satellite Image Analysis for Locating Prehistoric
Archaeological Sites in Alaska's Central Brooks Range.” Journal of Archaeological
Science: Reports 3: 80-89. doi:10.1016/j.jasrep.2015.05.022.
https://www.sciencedirect.com/science/article/pii/S2352409X15300171.
Lekson, S., T. Windes, J. Stein, and W. Judge. 1988. “The Chaco Canyon Community.”
Scientific American 259 (1): 100-109. http://www.jstor.org/stable/24989164.
Oswald, B. 2018. “Chaco Canyon.” World History Encyclopedia.
https://www.worldhistory.org/Chaco_Canyon/.
NPCA. 2023. “Spoiled Parks Chaco Culture National Historical Park.” National Park
Conservation Association. https://www.npca.org/case-studies/chaco-culture-nationalhistorical-park.
NPS. 2003. “Chaco Culture National Historical Park Site Text.”, last modified June 16, 2023
https://www.nps.gov/museum/exhibits/chcu/chcu_alltext.htm.
NPS. 2024. “History & Culture.” National Parks Service Chaco Canyon National Historical Park
New Mexico. last modified March 11, accessed 03/09/, 2024,
https://www.nps.gov/chcu/learn/historyculture/index.htm.
Rappaport, G. 2011. “Guide to the Frank H. H. Roberts Jr. Photographs in MS 4851.”
Smithsonian Online Virtual Archives. https://sova.si.edu/record/naa.photolot.4851
Snead, J. E. 2012. “Obliterated Itineraries: Pueblo Trails, Chaco Roads, and Archaeological
Knowledge. Highways, Byways, and Road Systems in the Pre-Modern World.”
Academia.
https://www.academia.edu/6297441/Obliterated_Itineraries_Pueblo_Trails_Chaco_Roads
_and_Archaeological_Knowledge.
Toeppen, J. “Chaco Canyon 3D Models.” Sketch fab., accessed 02/09/, 2024,
https://sketchfab.com/toeppen/collections/chaco-canyon-new-mexico05e31adccbac4dcab320ad9241666656.
UNESCO. 2024. “Chaco Culture.” UNESCO World Heritage Convention.,
https://whc.unesco.org/en/list/353/.
Vivian, R. G. 1997a. “Chacoan Roads: Function.” Kiva 63 (1): 35–67.
doi:10.1080/00231940.1997.11758346.
https://doi.org/10.1080/00231940.1997.11758346.
———. 1997b. “Chacoan Roads: Morphology.” Kiva 63 (1): 7–34.
doi:10.1080/00231940.1997.11758345.
https://doi.org/10.1080/00231940.1997.11758345.
Watson, A. 2012. “Craft, Subsistence, and Political Change: An Archaeological Investigation of
Power and Economy in Prehistoric Chaco Canyon, New Mexico, 850 to 1200 CE.”

81
Ph.D., University of Virginia. https://www.proquest.com/docview/929140708?pqorigsite=gscholar&fromopenview=true&sourcetype=Dissertations%20&%20Theses#
Witts, D. 2010. “The Remote Sensing of Ancestral Puebloan Sites.” Master’s Thesis,
Department of Geography, the State University of New York at Buffalo.
Wills, W. 2017. “Water Management and the Political Economy of Chaco Canyon during the
Bonito Phase (Ca. AD 850–1200).” Kiva 83 (4): 369-413.
doi:10.1080/00231940.2017.1343109. https://doi.org/10.1080/00231940.2017.1343109

Abstract (if available)

Abstract This thesis project focuses on using advanced remote sensing technology to create a comprehensive 3D model of Chaco Canyon, a significant US National Historical Park and UNESCO World Heritage Site located in New Mexico. There are two primary motivations for this research. The first motivation is that a contiguous 3D Visualization of Chaco Canyon that is available for public viewing does not yet exist. The other primary motivation is to explore an under-documented archaeological site within Chaco Canyon that is vulnerable to becoming fully deteriorated, according to the results of this research. The project is built upon three interrelated sub-research objectives. First, a supervised image classification is conducted on satellite imagery of Chaco Canyon, which is then utilized to identify new archaeological areas of interest based on the spectral signatures of the known ruins and roadways. Second, LiDAR data is used to investigate a concealed site, and then an LCP analysis is performed to model potential travel routes to nearby ruins from the concealed site. To conclude, a 3D visualization of Chaco Canyon is generated using LiDAR point cloud data and high-resolution satellite imagery. By employing cutting-edge remote sensing techniques and GIS methodologies, the product of this project is a contiguous 3D visualization of Chaco Canyon that displays the locations of the major ruins and illustrates the extent of the ancient roadway network. Additionally, this study seeks to support ongoing cultural heritage preservation efforts at Chaco Canyon. The findings will benefit the National Parks Service, associated tribes, conservation groups, and the broader academic and public communities by providing a complete 3D visualization that can be used for educational purposes, preservation efforts, informing public policy, and as a foundation for future archaeological research at Chaco Canyon.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Assessing modern conflict to monitor human rights with remote sensing: Russia's war in Ukraine

PDF

Classifying ash cloud attributes of Eyjafjallajökull volcano, Iceland, using satellite remote sensing

PDF

Satellite derived bathymetry in the Canadian Archipelago using multispectral and LiDAR space-based remote sensing

PDF

Assessing the use of normalized difference chlorophyll index to estimate chlorophyll-A concentrations using Landsat 5 TM and Landsat 8 OLI imagery in the Salton Sea, California

PDF

An examination of close-range photogrammetry and traditional cave survey methods for terrestrial and underwater caves for 3-dimensional mapping

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

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

Asset Metadata

Creator Penna, Paul T. (author)

Core Title From ruins to pixels: using remote sensing and GIS to analyze, document, and visualize archaeological sites and ancient roadways in Chaco Canyon, New Mexico

School College of Letters, Arts and Sciences

Degree Master of Science

Degree Program Geographic Information Science and Technology

Degree Conferral Date 2024-12

Publication Date 09/26/2024

Defense Date 08/29/2024

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

Tag 3D modeling,American Southwest,archaeology,LiDAR,OAI-PMH Harvest,remote sensing,satellite imagery

Format theses (aat)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Sedano, Elisabeth (committee chair), Dodd, Lynn (committee member), Qi, Yi (committee member)

Creator Email ppenna@usc.edu,ptpenna@icloud.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-oUC11399B9YS

Unique identifier UC11399B9YS

Identifier etd-PennaPaulT-13546.pdf (filename)

Legacy Identifier etd-PennaPaulT-13546

Document Type Thesis

Format theses (aat)

Rights Penna, Paul T.

Internet Media Type application/pdf

Type texts

Source 20240926-usctheses-batch-1214 (batch), 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 author, as the original true and official version of the work, but does not grant the reader permission to use the work if the desired use is covered by copyright. It is the author, as rights holder, who must provide use permission if such use is covered by copyright.

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

Repository Email cisadmin@lib.usc.edu