University of Southern California Dissertations and Theses

A comparison of urban land cover change: a study of Pasadena and Inglewood, California, 1992‐2011

(USC Thesis Other)

A comparison of urban land cover change: a study of Pasadena and Inglewood, California, 1992‐2011

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content
A COMPARISON OF URBAN LAND COVER CHANGE:
A STUDY OF PASADENA AND INGLEWOOD, CALIFORNIA, 1992 - 2011

by

Richard J. Crowther Jr.

A Thesis Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
MASTER OF SCIENCE
(GEOGRAPHIC INFORMATION SCIENCE AND TECHNOLOGY)

August 2015

Copyright 2015 Richard J. Crowther Jr.
ii

DEDICATION
I dedicate my thesis in honor of my grandparents, who I would not be here without their love and
support. I also dedicate my thesis to my loving parents who have supported me throughout
everything, my sister for all of her encouragement to graduate, and my godmother for all the
stress-relief phone calls. Last, but not least, this is for my cousins and friends for their support
and encouragement.

iii

ACKNOWLEDGEMENTS
I am also grateful to Daniel Warshawsky, assistant professor (teaching) at the Spatial Sciences
Institute. I am extremely thankful to him for his sincere and valuable guidance and
encouragement. Thank you for putting up with me.

iv

TABLE OF CONTENTS

DEDICATION ii

ACKNOWLEDGEMENTS iii

LIST OF TABLES vi

LIST OF FIGURES viii

LIST OF ABBREVIATIONS ix

ABSTRACT x

CHAPTER 1: INTRODUCTION 1
1.1 Landsat Imagery and Land Cover Change 1
1.2 Existing Research Gaps 2
1.3 An Outlook on Pasadena and Inglewood 4
1.3.1 Pasadena 5
1.3.2 Inglewood 7
1.4 Thesis Orientation 8
CHAPTER 2: RELATED WORK 10
2.1 Previously Completed Studies 10
2.2 Land Cover Change over the Past Century 13
CHAPTER 3: METHODOLOGY 17
3.1 Framework 17
3.2 Data Acquisition 18
3.3 Limitations of Imagery 19
3.4 Data Analysis 22

v

CHAPTER 4: RESULTS 30
4.1 Comparison of Land Cover, 1992 to 2001 30
4.1.1 1992 Classifications 31
4.1.2 2002 Classifications 34
4.1.3 1992-2001 Land Cover Percent Changes 37
4.2 Comparison of Land Cover, 2001 to 2011 39
4.2.1 2001 Classifications 39
4.2.2 2011 Classifications 42
4.2.3 2001-2011 Land Cover Percent Changes 46
4.3 Pasadena vs. Inglewood 47
CHAPTER 5: DISCUSSION & REVIEW 49
5.1 Observations and Limitations 49
5.2 Significance of Findings 50
5.3 Future Research 51
REFERENCES 53
vi

LIST OF TABLES
Table 1: Listing of Landsat satellites with specifications 20

Table 2: Listing of each sensor used on each of the Landsat satellites 21

Table 3: Class descriptions for NLCD 1992 classifications 26

Table 4: Listing of Anderson I classifications 26

Table 5a: The reclassification of the NLCD 1992 classes 27

Table 5b: The reclassification of the NLCD 2001 classes 27

Table 6: Class descriptions for NLCD 2001 and 2011 classifications 28

Table 7: Percentage of land cover by Anderson I class for Pasadena from 31
NLCD 1992 data

Table 8: Percentage of land cover by Anderson I class for Inglewood from 31
NLCD 1992 data

Table 9: Percentage of land cover by Anderson I class for Pasadena from 37
NLCD 2001 data

Table 10: Percentage of land cover by Anderson I class for Inglewood from 37
NLCD 2001 data

Table 11: Land cover percent change of Pasadena from 1992 to 2001 38
Table 12: Land cover percent change of Inglewood from 1992 to 2001 38
Table 13: Percentage of land cover by decision-tree class for Pasadena from 39
NLCD 2001 data

Table 14: Percentage of land cover by decision-tree class for Inglewood from 42
NLCD 2001 data

Table 15: Percentage of land cover by decision-tree class for Pasadena from 43
NLCD 2011 data

Table 16: Percentage of land cover by decision-tree class for Inglewood from 43
NLCD 2011 data

Table 17: Land cover percent change of Pasadena from 2001 to 2011 46

vii

Table 18: Land cover percent change of Pasadena from 2001 to 2011 47

viii

LIST OF FIGURES
Figure 1: Location Comparison Map of Pasadena and Inglewood 6
Figure 2: Analysis of spatio-temporal land cover change patterns in Pasadena and 18
Inglewood, California from 1992 to 2011

Figure 3: NLCD 2011 data of the study areas 23
Figure 4: Anderson I classification of land cover for Inglewood in 1992 32
Figure 5: Anderson I classification of land cover for Pasadena in 1992 33
Figure 6: Anderson I classification of land cover for Inglewood in 2001 35
Figure 7: Anderson I classification of land cover for Pasadena in 2001 36
Figure 8: Decision-tree classification of land cover for Inglewood in 2001 40
Figure 9: Decision-tree classification of land cover for Pasadena in 2001 41
Figure 10: Decision-tree classification of land cover for Inglewood in 2011 44
Figure 11: Decision-tree classification of land cover for Pasadena in 2011 45
Figure 12: Bar graph displaying the land cover change by classifications over the 48
twenty year period.

ix

LIST OF ABBREVIATIONS
NLCD National Land Cover Database
GIS Geographical Information Systems
SPOT Satellite Pour l'Observation de la Terre
MODIS Moderate Resolution Imaging Spectroradiometer
PHX-UGM Phoenix Urban Growth Model
MRLC Multi-Resolution Land Characteristics Consortium
MSS Multispectral Scanner
TM Thermal Mapper
ETM+ Enhanced Thermal Mapper Plus
OLI Operational Land Imager
TIRS Thermal Infrared Sensor
DEM Digital Elevation Model
CIR Color Infrared Composite
USGS United States Geological Survey
EPA Environmental Protection Agency
NOAA National Oceanic and Atmospheric Administration
USFS United States Forestry Service
BLM Bureau of Land Management
NASA National Space and Aeronautics Administration
NPS National Parks Service
USNFWS United States National Fish and Wildlife Service
NASS National Agriculture Statistics Service
x

ABSTRACT
Imagery and spatial data collected from different tools and satellite technologies have been used
to complete land cover change studies at the scale of cities, countries and continents. Different
methodologies have been used to complete these studies, dependent upon the technology and
information available to complete land cover change. In this thesis, urban land cover has been
analyzed by applying Landsat satellite imagery to spatial analysis as a way to examine land cover
changes in Pasadena, California and Inglewood, California from 1992 to 2011. The objective for
this study has been to review spatial data collected from Landsat data in order to understand
urban land cover change in each city. Spatial data collected from the National Land Cover
Database (NLCD) have been pre-processed with color infrared composite creation and image
classification tools to show land cover. Imagery from Landsat 4 has been used to help compare
land cover change from 1992 to 2001 since classifications of the NLCD were different in both
years. The resulting maps display the land cover changes over time from the effective application
of imagery analysis to complete a pattern of land cover change over the time of twenty years.
The study’s findings demonstrate that cities in the same metropolitan center can have similar
urban growth patterns even when they have geographically diverse landscapes. These findings
underscore the importance of understanding urban grown patterns when planning for urban.

1

CHAPTER 1: INTRODUCTION
Satellite imagery has enhanced the way the world can be viewed. This type of imagery can be
manipulated and run through spatial analysis classification tools to further understand our world.
Taking imagery from multiple decades and applying spatial analysis classification tools can
illuminate the changes that have occurred during that period of time. Understanding the urban
and population changes of the past can help the future of urban planning.
1.1 Landsat Imagery and Land Cover Change
To understand the structure, function, and dynamics of urban areas, it is necessary to integrate
both ecological and human progress, such as physical geographic changes to road maps or
population change, that result in land cover change. As a result of human activities, pervasive
ecological changes have occurred at local, regional, and global scales, such as the change in land
cover of natural landscapes to provide for human needs (Vitousek 1994, 1867-68). Urban
landscapes exhibit the most conspicuous spatial heterogeneity of all landscapes, and the spatial
form a city takes affects physical, ecological, and sociological processes (Picket et al., 1997;
Zipperer et al. 2000; Wu and David 2002). While the ecological and sociological effects of land
conversion for agricultural uses have been studied (Riebsame et al. 1994, 351), the effects of
land conversion for human habitation, or urbanization, are less understood (Pickett et al. 1997,
186). Urbanization is the general process of city growth; native land cover is appropriated for
industrial, commercial, residential, and other land uses associated with human demands. As
human population increases and as increasing proportions of people move to urban
environments, the number and size of urbanized areas will also increase globally (Simpson 1993;
Cohen 1995). The importance of these particularly human-dominated landscapes in controlling
global biospheric processes is particularly important (Jenerette and Wu 2001, 611). The patterns
2

and processes of urbanization should be integrated if the ecology of cities is to be fully
understood (Foresman et al. 1997; Wu and David 2002).
A distinction between what is land cover and what is land use will help understand the
focus of this study. Land use describes use of the land surface by humans, usually in an
economic context, as in residential, commercial, agricultural (Campbell and Wayne 2011, 585).
However land use can be hard to see, except under close examination. The visible features of the
Earth’s surface, such as vegetation cover, natural and modified by humans, structures, all fall
comprise land cover. This study looks at the land cover in the imagery of the study areas.
Understanding the patterns and processes of land cover change over time has become
easier to understand with the ability to interpolate satellite imagery to classify land cover types
and the manner in the land cover has changed by comparing different periods of imagery with
geographical information systems (GIS) software. The urban footprint can be seen through
different remote sensing tools to provide imagery with real time conditions. In the case of this
study, imagery of Pasadena and Inglewood, California has been obtained from Landsat 4
Thematic Mapper satellite sensors at the USGS GLOVIS website. These data have been
analyzed with ArcMap 10.1, utilizing spatial analysis and image classification tools to achieve
land cover types that are compared with NLCD data.
1.2 Existing Research Gaps
This study includes two non-coterminous, large populous and geographically different cities in
Southern California; however, previous studies have covered much broader study areas. Some
previous studies include provinces/states or even entire countries (Deng et al., 2009; Tian et al.
2005; Jenerette and Wu 2001), while others cover large metropolitan regions (Luck and Wu
2002; Berling-Wolff and Wu 2004; Shenghe, Prieler and Xiubin 2002). The ability to use large
3

amounts of data enables for more work to correct inaccuracy due to the multitude of data
involved. The ability of this study of Pasadena and Inglewood is to focus on two determinate
locales within the same metropolitan area and begin the analysis with accuracy as the image
classification work begins. However, this is not always a feature as there is usually some
leniency with data accuracy that covers large areas.
Secondly, working with NLCD land cover data from 1992 and data from more recent
years is not a simple task. The land cover classification utilized for the NLCD 1992 is based off
the Anderson method for classification from the late 1970s. This method worked well with the
satellite imagery from the Landsat 4 satellite in the early 1990s, however, did not work well with
the more technologically advanced imagery that occurred thereafter, and a new See5 decision-
tree classification method. This method is user friendly, and more importantly, allowed for
better classification of urban elements based on imperviousness (Fry et al. 2008, 2). Using a
broader classification system that is also based on the Anderson classifications method, the
NLCD 1992 land cover data can be compared with land cover data gathered in the more recent
years.
Another factor to keep in mind is time, as in the period of time in which the study is
completed. There have been studies that span over centuries (Meiyappan and Jain 2012, 122;
Aspinall 2004, 91), and yet others that only consider a decade of land cover change (Tian et al.
2005; Shenghe, Prieler and Xiubin 2002). There does not seem to be a corresponding time limit
when it comes to these studies other than the limit to the data available to complete spatial
analysis with satellite imagery. This thesis is based on the timeline of NLCD data from earliest
available to the most recent.
4

The last point gathered from researching land cover change with GIS is that there has not
been one single way to complete the analysis. Some studies use modeling to complete land
cover change over time using a variety of data sources (Berling-Wolff and Wu 2004; Jenerette
and Wu 2001; Aspinall 2004, 91). On the other hand, there are other studies that used satellite
imagery to complete the urban land cover change analysis, but have different systems for the
data. Imagery from the Landsat family, Satellite Pour l’Observation de la Terre,

translated into
English as Satellite for observation of Earth (SPOT), or Moderate-resolution Imaging
Spectroradiometer (MODIS) on either the Terra or Aqua satellites (it was not stated by
Meiyappan and Jain, 2012, on which was used) have been used with spatial analysis tools to
complete land cover change studies.
The case for this study is to utilize a sound method of urban land cover change analysis
based on image quality, data availability, and data manipulation. Previous studies have been
completed to gain knowledge of certain areas and data was collected for those places. This study
will focus on determinate places and a time period with the best data available to use and
improve accuracy.
1.3 An Outlook on Pasadena and Inglewood
Understanding the past of these two cities ensures that the analysis is probably situated and
ensures that we know to look for, i.e. places of constant change, parts of the city that will never
change due to laws protecting open land or historic areas, etc. Without knowing what happened
before the use of satellite imagery to examine land cover, it is hard to understand what and why
the changes that are found through the use of this imagery might have occurred. The previous
map shows the location of the study areas in Los Angeles County, with the inset showing a
5

closer geographical relationship. The following briefly discusses the past of the two cities under
study.
1.3.1 Pasadena
Historically, the land of Pasadena transitioned from agricultural farmland in the city’s beginnings
to a copy of an Industrial Revolution age city in the start of the twentieth century to the now
technological and cultural focused urban lifestyle Pasadena is recognized as. The era of
agricultural prosperity in Pasadena predates the technology to be used in analysis, but one can
assume that urbanization drastically changed the look of the southern California city.
People were migrating west for new opportunities and industry was creating new jobs to be
filled. Pasadena’s population at the time of the Industrial Revolution in the 1920s was nearly ten
times of what it was at the time of its incorporation in 1886 (Los Angeles Almanac 2014).
The population of the city during the twenty years of this study has grown from 131,591 in 1990
to 133,936 in 2000 and up to 137,122 in 2010 according to the United States Census. The
majority of Pasadena is of white ethnicity, followed by Hispanic/Latino, Asian, and
Black/African-American. The high school graduation rate is 85.6%, and about half of the adult
population hold a Bachelor’s degree or higher. As of 2010, there are 55,110 households in
Pasadena, with an average of 2.43 persons per household. In 1990, the number of households
was lower at 50,199 with an average of 2.53 persons per household. The city consumes 22.97
square miles of Los Angeles County. This leads to 5,969.6 persons per square mile, as of 2010.
Due to significant changes in land cover over the last few decades and great data
availability, Pasadena makes for an interesting case study. The city is geographically diverse
with both urban and natural landscape features. In addition, Pasadena is a good choice for this
6

study since analysis can be cross-checked for study accuracy with its real physical locations in
Southern California.
Another reason for the choice of Pasadena as a study area is the understanding of how the
city has been planned out. Since Pasadena’s beginning, the layout of streets, neighborhoods, and
open space has been carefully architected. The current land use policies are now focused on
density issues and creating a sense of place, great neighborhoods, gardens, plazas, parks, and
trees (Land Use and Mobility Element Policies 2014). The latter is especially important since the
city is known as Tree City USA, as stated by a sign in Pasadena.

Figure 1. Location Comparison Map of Pasadena to Inglewood

7

1.3.2 Inglewood
With its past in agriculture as well, Inglewood made its mark in barley and poultry production.
The city was incorporated in 1908 with a population of 1200. A small number by today’s
perspective, but in the early 1920s, Inglewood was known as the fastest growing city in the
United States, mainly from visitors seeking earthquake damage and stayed for the coastal
environment. Also in the 1920s, Inglewood became to be known as the chinchilla capital of the
World as the locale was easy to import to from Peru.
World War II turned Inglewood from an agricultural hub into an urban community built
by defense industries. During the 1960s and 1970s, the city continued to grow into a
metropolitan city, embracing both its residential and commercial communities, and supporting
sports with the Hollywood Park racetrack and having a few Olympians who grew up in the city.
With its ease of accessibility to major freeways and Los Angeles International Airport, it made
for an attractive business environment and ideal location for air freight business. The skyline
began to change as more tall municipal and commercial buildings hovered over the historical
places the citizens continue to protect and praise (City of Inglewood, 2015). Inglewood, unlike
Pasadena, is not geographically diverse and is highly urbanized. This opposition makes
Inglewood a good choice to compare against Pasadena in this case study.
Inglewood was chosen for its homogeneous physical landscape and for having an uneven
population growth rate, to offset that of Pasadena. Inglewood is also located in Los Angeles
County, but as its own entity like Pasadena. As seen in Figure 1, Inglewood and Pasadena are
only separated by the city of Los Angeles. The population of Inglewood only grew by seventy
from 1990 to 2010, as it went from 109,602 to 109,673. However, in 2000, the population
spiked at 119,580 according to the U.S. Census. This means that there was growth, but an equal
8

decline within those twenty years. The number of households reflected the same fluctuation,
beginning with 36,102 households in 1990, rising to 36,805 in 2000, and declining to 36,461 in
2010. Inglewood comprises 9.07 square miles of land south of Los Angeles. Using the
population totals for 2010, this leads to 12,094.5 people per square mile.
The majority ethnicity has also changed in those two decades. In 1990, the highest
percentage of the population was fifty-two percent Black/African-American. While
Black/African-American remained the majority in 2000, it declined with a rising Hispanic/Latino
population. In 2010, the majority became Hispanic/Latino at 50.56%. All the meanwhile, the
White percentage climbed higher from 17.47% in 1990 to 23.31% in 2010. Also, according to
the U.S. Census, the high school graduation rate increased from 19.8% in the early nineties to
71.8% by 2010.
Inglewood also has a General Plan for the city’s land use, mostly focusing on
commercial, industrial, and transportation sections, with housing, open space, and other uses also
under control, however, not as detailed as that of Pasadena. Having two different study areas
that are geographically and socio-economically different enables this study to demonstrate the
ability of land cover classification with different conditions involves, and to understand any
demographic characteristic relationships.
1.4 Thesis Orientation
The objective for this study is to analyze land cover data for Pasadena and Inglewood, California
from 1992, 2001, and 2011 to cover a twenty year timeline of urban land cover change in the
greater Los Angeles metropolitan area to gain a better understanding of past land cover changes
to prepare for the future in urban planning. This includes data acquisition, image classification,
and urban land index comparison of NLCD data for 1992 and 2001, to be compared with NLCD
9

data for the changes from 2001 to 2011. This allows the information to reflect future transitions
within the city and help develop urban planning ideas for the community.
This thesis is divided into five chapters. Chapter Two reviews the development of the
use of Landsat imagery, and satellite imagery in general, in GIS for land cover change analysis,
along with a history of the study of land cover change in the past century. This chapter describes
the different methods for spatial analysis including satellite imagery for understanding patterns
of land cover change. Chapter Three describes the study area, data sources, and the methods for
the image classification of the Landsat imagery. The conceptual framework is explained along
with the expected outcomes. Chapter Four reviews and explains the outcomes of the study.
Chapter Five discusses the impact of the findings and their contribution to existing research on
urban land cover change over time.

10

CHAPTER 2: RELATED WORK
Geospatial analysis for urban land cover change over time has benefited from the use of GIS
tools to analyze data and display results. Completing a review of previous studies gives an idea
of what can occur in this study and what the results could be. This review also presents an idea
of the obstacles to further spatial analysis of the data. Also, since this study will use previously
completed image classification for land cover types from satellite imagery of Southern California
to compare the land cover change of two cities to understand the urban growth in the Los
Angeles metropolitan region, a review of how studying land cover has evolved over time will be
conducted.
2.1 Previously Completed Studies
The following studies discuss the use of satellite imagery in land cover change in different parts
of the United States and the world. The ideologies behind these studies helped guide in choosing
the methodology for this study on land cover change.
This first study was the primary influence for the methodology of the study of Pasadena
and Inglewood. Guangjin Tian and co-authors studied the spatio-temporal characteristics of
urban expansion in China using Landsat TM imagery from 1990/1991, 1995/1996, and
1999/2000. These images interpreted to a 1:10000 vector land cover dataset helped to clarify
where the fast urban growth occurred regionally throughout the entire country of China. Tian
and co-authors calculated the urban land percentage and urban land expansion index of every 1
km
2
cell that covered China. The previous two terms are furthered explained as they are used in
the method of completing the study. Their result was land cover dynamic changes reflect the
strong impacts of economic growth regions and urban development policies.
11

Staying with the study area of China, Jin S. Deng and his co-authors focused in the
Zhejiang province from 1996 to 2006. Their theory was the “analysis of spatio-temporal
characteristics of land cover change was essential for understanding and assessing ecological
consequences of urbanization (Deng et al. 2009, 187). They utilized SPOT imagery, along with
spatial metrics derived from FRAGSTATS software to find that the rapid urbanization process
included large amounts of land cover change and urban growth at high rates, changing the look
of the landscape. Deng and his cohorts’ methodology was based on pre-processing the data,
enhancing and extracting land cover change information, and assessing the accuracy of the
results. Their results confirmed the effectiveness of the combined method of using remote
sensing and metrics to define spatio-temporal features to further define land cover patterns across
time periods.
The third study is by Liu Shenghe, Sylvia Prieler, and Li Xiubin based on the urban
growth centers of Beijing, China. They digitized land cover maps from 1982, 1992, and 1997 to
create data to be used with spatial analysis tools. This study could have been completed with the
use of satellite imagery to increase the accuracy of the data since digitizing decreases accuracy.
G. Darrel Jenerette and Jianguo Wu’s land use methodology created a boundary line in
the Central Arizona/Phoenix region and gained an understanding of what was urban, desert,
agriculture, or other land cover classifications within that boundary. Their methodology
consisted of creating landscape raster maps with 250 m
2
pixels, classifying the majority rule
criterion of their design, and using a multiple of digressions, such as edge density, fractal
dimension, and contagion/landscape configuration. This was completed to examine the
relationships between potential socioeconomic drivers and environmental constraints, land cover
pattern correlation of population size through population growth model and correlation of
12

topography though estimated mean slope of each land cover class by overlaying a digital
elevation model (DEM) onto each land cover map. A probabilistic cellular automata simulation
model was used to complete a temporal sequence of land cover change pattern. This study was
the inspiration to gain an understanding of the affect population has on the change in land cover
in Pasadena and Inglewood.
Other studies conducted about the Phoenix area have been done by Matthew Luck and
Jianguo Wu, and Sheryl Berling-Wolff and Jianguo Wu. Luck and Wu converted a vector
dataset to a raster data set with a 50 in
2
pixel after resetting classifications from a 1995 Maricopa
land cover dataset. Landscape metrics and a gradient transect were applied to the raster dataset
to attempt to identify land cover type spatial signatures. Their approach was not land cover
change over time, but land cover change over area and if there are any apparent spatial patterns
of urbanization. Berling-Wolff and Wu also quantified spatial patterns of urbanization, along
with temporal patterns of urbanization with dynamic modelling and spatial analysis. They
developed the Phoenix Urban Growth Model (PHX-UGM), a spatially explicit urban landscape
model, from the Human-Induced Land Transformations (HILT) model used in the San Francisco
Bay area. Selected growth rules and new rules were added to appropriately identify ecological
and social features of the Phoenix study area in the modelling. The methodology for this study
would not work with Landsat imagery due to the best results at a 120 – 450m spatial resolution,
and Landsat is 30m resolution.
Continuing with the use of model analysis for land cover change, Richard Aspinall links
model selection and multi-model inference with empirical models and GIS for the Gallatin
Valley in Montana between 1860 and 2000. He developed empirical models based on data for
case studies of land cover change. The choice between multiple alternate empirical models that
13

potentially can be developed for a given dataset was made for the best relationship of empirical
modelling to drivers of land cover change.
The choice to use the methodology based on satellite imagery spatial analysis tools in
ArcMap was used due to the complexity of the use of modelling to complete the study in
Pasadena and Inglewood. A. Veldkamp and E.F. Lambin agree that land cover change models
add to the complexity that is land cover systems. They add that land cover modelling can offer
the possibility of testing the sensitivity of land cover patterns when variables are changed and
these models can allow for the testing of the stability of linked social and ecological systems
throughout the land cover systems. This also furthers the Pasadena/Inglewood study context of
urban land cover change with population statistics. People make the changes; urban structure
just does not change by itself. The modelling methodology is reviewed to demonstrate what can
be done in understanding urban land cover change.
2.2 Land Cover Change over the Past Century
Land cover has been monitored since medieval times through cadastral maps of villages and
kingdoms to monitor crops and people. As time and technology progressed, the way the world is
looked at has changed. Aerial photography and satellites have given large area monitoring and
images a huge leap forward in how land can be managed.
Cadastral maps in Sweden from the seventeenth and eighteenth centuries kept record of
land use, ownership, physical features, yield, and quality of hay (Cousins 2001, 41). These maps
demonstrate the importance of understanding the land for the suit of use. These maps were later
used in a modern study as a source of information on present day ecological patterns. Sarah
A.O. Cousins used local warping to rectify the cadastral maps to determine general trends in land
use over three hundred years in Sweden to underestimate the full dynamics in land use change.
14

The next step after hand-drawn maps for understanding land cover is aerial imagery.
Aerial photos represent a very useful tool for the identification of past and present ground
features on large areas since photography has been available for over 150 years (Gennaretti
2011, 542). Aerial images can be used as basemaps or the imagery can be interpreted to
categorize or assign attributes to surface features. Specialized cameras can minimize distortion
and maximize image quality have been attached to aircraft, more recently to specialty aircraft
designed for photographic mapping projects (Bolstad 2012, 228).
Image interpreters use the fundamentals of land cover classification to delineate separate
land cover classes in an organized and systematic manner. This process has become
computerized and is now completed much faster when first devised. A study completed by
Fabio Gennaretti and his co-authors used some of the first aerial photos in France by
photographer Nadar from 1858 while aloft in an aerostat, an unpowered balloon or airship filled
with gas. They explained different procedures in which land cover mapping could be completed
from those images. The most common procedure was based on photo interpretation and manual
digitization of the land cover classes represented. However, this procedure proved time
consuming for large study areas and was subject to individual interpretation. Another procedure
for land cover maps to be obtained was by digital processing through the allocation of pixels into
finite number of classes or categories on the base of their spectral values. This approach used
images with few spectral bands and small pixel size that lead to classification errors.
A third procedure was object oriented which represented an updated method for the
extraction of information from aerial photos for land cover. This process involved the
segmentation of the initial images allowing for the extraction of vectorial polygons called
objects, marking off homogenous areas for spectral signature, shape, position, and texture
15

(Gennaretti 2011, 544). This process can be used to develop an automated procedure permitting
a computerized classification of land cover, which occurred for the process of Landsat imagery
to be used in the NLCD databases.
During the 1990s, digital aerial cameras were developed and proved more advantageous
to film cameras previously used since the 1920s. Digital cameras allowed for more flexibility,
stability, easier planning and execution, and direct digital output (Bolstad 2012, 230). Giles M.
Foody considered the accuracy assessment of imagery in the digital age. The following are four
historical stages of how accuracy of interpretation and analysis of imagery Foody listed. The
first stage was basic visual appraisal of the derived map – basically if the resulting map looked
good or not. The second stage was characterized by an attempt to quantify accuracy more
objectively. The accuracy assessment was based on comparisons of the areal extent of the
classes in the derived thematic map relative to their extent in some ground or other reference data
set. The limitation of this process was representation of the class proportions was correct,
however, in the wrong location. The third stage involved the derivation of accuracy metrics that
were based on a comparison of the class labels in the thematic map and ground data for a set of
locations. The fourth stage was a refinement of the third stage with a greater use of the
information on the correspondence of the predicted map labels to those observed on the ground
(Foody 2002, 190). These steps show the progress that has been made with aerial photos in the
past decades.
In the 1960s and 1970s, the development of satellites allowed for a new point of view for
imagery, about 400 miles away in space. Many satellites and tools collect imagery and data
about the Earth’s surface and more every day. The imagery covers more area and also better
quality over that of aerial imagery due to the distance and positioning. Satellite imagery is also
16

digital, allowing for computerized manipulation and classification. Satellite imagery and digital
aerial imagery are made of pixels that can be classified by spectral bands and comparing them
against one another. Satellite sensors began with four spectral bands and an infrared band and
now have sensors that use 8 bands, to be further discussed in the next chapter.

17

CHAPTER 3: METHODOLOGY
This study’s main examination of the physical urban characteristics, such as streets and
highways, commercial and residential buildings, and parks and forests, helps to delineate the
spatial-temporal land cover change patterns over the past twenty years. Then to identify and to
describe the spatial patterns that occur over time to help understand how a city’s urban footprint
grows.
3.1 Framework
The analysis of this study was encouraged by the procedure used in a study by Tian, Liu, Xie,
Yang, Zhuang, and Niu, in which each period of Landsat imagery is classified with the same
method and then compared to the next chronological period. The extent of the methodology of
their study is increased in this study from a five-year period that constituted their study to a
comparison of imagery from three decades. Yet, that study included creating a mosaic of many
images to include the whole country of China. This study excludes NLCD imagery data from
1992, 2001, and 2011 from the dataset that includes data from outside the study areas. The
imagery is applied in ArcMap so the work is stored in a file geodatabase and utilized in the
image classification tools. The resulting maps with the percentage of urban land cover by class
are compared to the following decade with the Urban Land Expansion Index. The work flow
chart in Figure 2 provides a visual understanding of this study’s procedure of examining land
cover change over time.
The land cover change will be compared with two different classification systems that the
NLCD was classified with. The NLCD 1992 data and the NLCD 2001 data will be compared
based on Anderson I classifications to merge the two different classifications used and the NLCD
18

2001 and NCLD 2011 data will be compared based on the decision-tree classifications used since
2001.

Figure 2. Analysis of spatio-temporal land cover change patterns in Pasadena and
Inglewood, California from 1992 to 2011.

3.2 Data Acquisition
Two sets of information are needed for the land cover change examination, NLCD data and
Landsat imagery. The NLCD data for all three years was obtained from Multi-Resolution Land
Characteristics Consortium (MRLC) website, www.mrlc.gov. The data available on this website
provides pre-classified land cover imagery for the United States.
For an accurate comparison, the land cover information for 1992 and 2001 had to fine an
equal ground. This allowed for the difference in land cover classification to be put into a broad
19

perspective over the study areas for this time period. The same land cover classification method
was used for 2001 and 2011 so there is no need for reclassification.
Comparing the land cover change from each year of imagery constitutes of Urban Land
Percentage (PU) which is generated from the image classification analysis distinguishing what
percentage of the total area (TL) is urban land cover (UL), shown in below.
PU = UL/TL x 100%
The PU is graded into a system that will show the significance of the land cover
classifications. The idea differentiates the urban developed areas and natural places such as
parks and forests. Parks are considered part of urban growth as Pasadena has prided itself on its
inclusion of parks and open space in the urban environment.
Next, the Urban Land Expansion Index (SI) is compared to the percentage of urban land
cover from each year to the next chronologically. Achieving the SI gives the percentage of the
land cover change between the two years being analyzed. The two years of UL are represented
in the following equation by ULj and ULi respectably.
SI = ULj – ULi/TL x 100%
The SI is graded into different classes showing the different degrees of land cover that
has occurred. The graded is also known after the completion of the image classification for 1992
and 2001. Dynamic urban change patterns can be determined by the SI, and thusly mapped.
3.3 Origins of the Imagery
As previously mentioned, satellite imagery did not exist previously to the 1950s, and with
Landsat satellites, not until the 1970s. The original intent of this study was to analyze urban land
cover change of the cities from earlier years, but it not feasible due to the lack of the proper data
for comparison.
20

The next limitation is the pixel count of the Landsat imagery from the various tools
aboard the satellites. Landsat satellites have carried two main tools aboard, a multispectral
scanner (MSS) and a thematic mapper (TM), in various forms. As technology advanced, so did
the Landsat satellites and the tools aboard them. The Landsat 4 satellite carried a multi-spectral
scanner and thematic mapper, in their earliest forms. The combination of these tools obtained
imagery of a 120 meter scale. The next generation, Landsat 5, included the same equipment, but
with better advancement to obtain 30 meter resolution. Landsat 7 incorporated a MSS, along
with and Enhanced Thematic Mapper Plus (ETM+). The latest in the Landsat lineage is the
Landsat 8, which comprises of an Operational Land Imager (OLI) and a Thermal InfraRed
Sensor (TIRS). The Landsat satellites and sensors that were utilized in the creation the land
cover datasets are Landsat 5 for the 1992 dataset and Landsat 7 for the 2001 and 2011 datasets.
Table 1 describes the facts about each of the satellites and Table 2 breaks down the two sensors
aboard each satellite.

Table 1. Listing of Landsat satellites with specifications
Specifications
Satellites
Landsat 5 Landsat 7
Launch Date March 1 ,1984 April 15, 1999
Sensors TM, MSS operational
Status MSS disabled ETM+
Altitude 704 km 705 km
Inclination 98.2 98.2
Orbit polar, sun-synchronous polar, sun-synchronous
Equatorial Crossing
Time nominally 9:45 am (15 min) nominally 10 am (15 min)
Period of Revolution 99 minutes (`14.5 orbits per day) 99 min (`14.5 orbits per day)
Repeat Coverage 16 days 16 days
Terminated June 5, 2013 Still Active
Duration 29 years, 3 months, 4 days 16 years, 1 month*
* reflects duration at time of table preparation
Source: http://landsat.gsfc.nasa.gov
21

Table 2. Listing of each sensor used on each of the Landsat satellites
Specifications
Satellite Sensors
Thermal Mapper
(TM)
Enhanced Thermal Mapper Plus
(ETM+)
Satellite Landsat 4 & 5 Landsat 7
Sensor Type opto-mechanical opto-mechanical
Spatial Resolution 30 m 30 m, thermal - 60 km, pan - 15 m
Spectral Range 0.45 - 12.5 um 0.45 - 12.5 um
Number of Bands 7 8
Temporal
Resolution 16 days 16 days
Image Size 185 km by 172 km 183 km by 170 km
Swath 185 km 183
Programmable Yes Yes
Source: http://landsat.gsfc.nasa.gov

The imagery from Landsat 5 Thematic Mapper was used to derive the data for the 1992
dataset by the MRLC (Vogelmann et al 2001, 651). Beginning in 1993, the 1992 prototype took
five years to map, with the 30 m resolution creating about nine billion pixels over the
conterminous United States. At this time, the MRLC consisted of four government groups: the
U.S Geological Survey (USGS), the Environmental Protection Agency (EPA), the National
Oceanic and Atmospheric Administration (NOAA), and the U.S. Forestry Service (USFS). This
consortium produced the product under the Anderson Level II classification. This classification
system proved to be inappropriate with the development of the 2001 database.
The MRLC grew after the success of the land cover classification of the 1992 dataset
with the addition of six more members: the Bureau of Land Management (BLM), the National
Aeronautics and Space Administration (NASA), the National Parks Service (NPS), the U.S. Fish
and Wildlife Service (USNFWS), the National Agricultural Statistics Service (NASS), and the
U.S. Army Corps of Engineers (Homer et al 2004, 832). All of these government groups had
22

sections to monitor within the databases. The NASS focused on land for agricultural use, the
NOAA monitored the coast lines, and the NFS managed tree canopy cover. The newly-enlarged
consortium had now the tools aboard the Landsat 7 satellite and needed a new way to classify
land cover. The Landsat 7 image acquisition included multi-temporal data processed to standard
procedures for three dates per path or row to represent three seasons for the United States and
Puerto Rico. Ancillary data, including 30 m DEM, slope, aspect, and positional index,
accompanied the imagery of the 88 mapping zones and a decision tree classification to create the
NLCD 2001. The process was repeated for the NLCD 2006 and NLCD 2011.
3.4 Data Analysis
This section describes the methods used to complete the image pre-processing and analysis to
achieve the Urban Land PUs and SIs to determine the land cover change. The science behind the
following process is based on the methodology of Tian et al.
The use of three different geodatabases (one for each year) allows for the data to be
organized throughout the processing and comparison analysis. The NLCD data from 2011 was
the first data to be analyzed. When working with satellite imagery, multiple images are gathered
from the different wavelength bands emitted from the Landsat 8 OLI remote sensing tool.
However, only three select bands are needed to properly analyze the data for land cover change,
Bands 5, 4, and 3 (in this order) to complete a color infrared (CIR) composite image to complete
imagery classification. The process of creating a CIR composite image involves the conversion
of multi-band raster imagery into a single-band raster with a number of categorical classes that
can be related to different types of land cover. To achieve the proper land cover classifications
to show the different types of urban elements of the land cover, training samples must be taken to
create a signature file to be used with a classification tool so this tool can understand what the
23

classifications are for analysis. This was then compared to the NLCD dataset for accuracy. All
of the land cover classification was completed by the MRLC and used in this study for accuracy.
The NLCD data was focused into the greater Los Angeles area to narrow on the two
study areas. The city boundary shapefile obtained from the Los Angeles County Geoportal
allowed for the use of city outlines to highlight the study areas. This boundary shapefile was
queried from the main shapefile that contained the political boundaries of all the municipalities
in Los Angeles County. Pasadena and Inglewood are the focus of this study, therefore there is
no advantage to completing a comparison of the whole image that covered the entire United
States, and the NLCD data was clipped to only include the two cities that are represented in
Figure 3.

Figure 3. NLCD 2011 data of the study areas

24

For this study, there are two main categories from the MRLC land cover classification,
developed areas and natural land cover. In the first comparison of NLCD 1992 to NCLD 2001,
the reclassification will create a class that covers all urban features, as well as one for all forested
areas, grassland and shrubland, agriculture, barren land, open water, and wetland. Originally in
the NLCD 1992 data, the first category encompasses all man-made land cover at four levels of
intensities of urban development that has occurred. In the case of the NLCD 1992 data, the
classes are Low Intensity Residential areas that are 30% to 80% impervious surfaces with some
vegetation, High Intensity Residential areas that are 80% to 100% impervious with some
vegetation, Commercial/Industrial/Transportation areas that are 100% impervious with no
vegetation, and Urban/Recreational Grasses that are semi-impervious areas with vegetation
like sport fields, golf courses, and parks (Vogelmann et al. 2001, 651-63). However, for an
accurate comparison, the classes will be grouped into the broader Urban class under the
Anderson I classification. The NLCD 2001 land cover data is also reclassified with the Anderson
I style for the comparison, but using different original classes.
For the NLCD 2001 and 2011, the first class is Developed Open Space areas with a
mixture of some constructed materials, but mostly vegetation in the form of lawn grasses.
Impervious surfaces account for less than 20% of total cover. These areas most commonly
include large-lot single-family housing units, parks, golf courses, and vegetation planted in
developed settings for recreation, erosion control, or aesthetic purposes. Next, Developed Low
Intensity areas contain a mixture of constructed materials and vegetation. Impervious surfaces
account for 20% to 49% percent of total cover. Developed Medium Intensity areas have a
mixture of constructed materials and vegetation. Impervious surfaces account for 50% to 79% of
the total cover. These areas most commonly include single-family housing units. Finally, the
25

Developed High Intensity class contains highly developed areas where people reside or work in
high numbers. Examples include apartment complexes, row houses, commercial and industrial
properties. Impervious surfaces account for 80% to 100% of the total cover. These four classes
are the main focal point for the urban land cover change in the study areas for the two cities are
mostly urban and contain little or none natural features.
The second category of classes focuses on the natural environment. Classes contain land
cover numbers for open water, vegetation, agriculture, and wetlands. Thinking of the land cover
of Southern California, one does not think of natural land cover being a large focal point, but
these classes are still important for comparison for urbanization usually lowers the percentage of
these classes occurring from decade to decade. As mentioned before, Pasadena comprises of
many large parks, plus forested areas in the mountainous sections of the city, so the main use of
this category is represented in the land cover changes of Pasadena.
A listing of all of the classes from the NLCD that are relevant in this study can be seen in
Table 3 for the 1992 classification and Table 6 for the 2001 and 2011 classifications. Tables 5a
and 5b displays the reclassifications for the Anderson I method for the 1992 to 2001 comparison.
Each of the classes has a number of occurrences in the image; therefore the percentage of each
class can be concluded from this. The SI formula is utilized to understand each individual
developed class land cover percentage and a larger look at the natural land cover.
However, to be used in this study, the classes have to be changed to the Anderson I
classification system to accurately compare 1992 to 2001 because the current classification
systems used for the NLCD 1992 and 2001 are not the same. The following table shows the
Anderson I classes for which the classes from the NLCD 1992 and 2001 merge into in Table 5a
and 5b.
26

Table 3. Class descriptions for NLCD 1992 classifications

Source: Data adapted from MRLC (2015)

Table 4. Listing of Anderson I classifications
Class
Number
Class Name
1 Open Water
2 Urban
3 Barren
4 Forest
5 Grassland/Shrub
6 Agriculture
7 Wetlands
Source: Data adapted from MRLC (2015)
Class Description
11 Open Water
21 Low Intensity Residential (Impervious Surface 30-80%)
22 High Intensity Residential (Impervious Surfaces 80-100%)
23 Commerical/Industrial.Transportation (100% Impervious)
31 Bare Rock/Sand/Clay
32 Quarries/Strip Mines/Gravel Pits
41 Deciduous Forest
42 Evergreen Forest
43 Mixed Forest
51 Shrubland (area dominated by shrubs)
61 Orchards/Vineyards/Other
71
Grassland/Herbaceous (unmaintained, dominayed by
herbaceous vegetation
81 Pasture/Hay (land intended for livestock grazing)
82 Row Crops (annual crop production)
83
Small Grains (areas used for the production of crops such as
wheat, barley, oats, and rice)
84
Fallow (areas used for the production of crops that do not
exhibit visible vegetation)
85 Urban/Recreational Grasses (parks, lawns, golf courses)
91
Woody Wetlands (areas where forest or shrubland
vegetation accounts for greater than 20% of vegetative cover
and the soil or substrate is periodically saturated with or
covered with water)
92
Emergent Herbaceous Wetlands (areas where perennial
herbaceous vegetation accounts for greater than 80% of
vegetative cover and the soil or substrate is periodically
saturated with or covered with water)
27

Table 5a. The reclassification of the NLCD 1992 classes.
NLCD 1992 Classification Number
Anderson I Classification
Number
Open Water Open Water
Low Intensity Residential Urban
High Intensity Residential Urban
Commercial/Industrial/Transportation Urban
Bare Rock/Sand/Clay Barren Land
Quarries/Strip Mines/Gravel Pits Barren Land
Deciduous Forest Forest
Evergreen Forest Forest
Mixed Forest Forest
Shrubland Grassland/Shrub
Orchards/Vineyards/other Agriculture
Grassland/Herbaceous Grassland/Shrub
Pasture/Hay Agriculture
Row Crops Agriculture
Small Grains Agriculture
Fallow Agriculture
Urban/Recreational Grasses Urban
Woody Wetlands Wetlands
Emergent Herbaceous Wetlands Wetlands

Table 5b. The reclassification of the NLCD 2001 classes.
NLCD 1992 Classification Number
Anderson I
Classification
Number
Open Water Open Water
Developed Open Space Urban
Developed Low Intensity Urban
Developed Medium Intensity Urban
Developed High Intensity Urban
Barren Land Barren Land
Deciduous Forest Forest
Evergreen Forest Forest
Mixed Forest Forest
Shrub/Scrub Grassland/Shrub
Grassland/Herbaceous Grassland/Shrub
Pasture/Hay Agriculture
Cultivated Crops Agriculture
Woody Wetlands Wetlands
Emergent Herbaceous Wetlands Wetlands
28

As stated earlier, the NLCD data for 2001 is sufficient to be accurately examined against
2011 since the classification methods were the same, but not accurate for a comparison with the
1992 data. Therefore, the data from 2001 has to be worked with twice, once like the methods in
2011 and then a more complex, time consuming cross-examination with Landsat imagery to
compare with the 1992 data. The MRLC has already completed a product that highlights areas
of potential change between 1992 and 2001, titled Retrofit LCC Data, which is used as a guide in
the land cover comparison of 1992 to 2001.

Table 6. Class descriptions for NLCD 2001 and 2011 classifications

Source: Data adapted from MRLC (2015)
Class Description
11 Open Water
21 Developed Open Space (Impervious Surface -20%)
22 Developed Low Intensity (Impervious Surfaces 20-49%)
23 Developed Medium Intensity (Impervious Surfaces 50-79%)
24 Developed High Intensity (Impervious Surface +80%)
31 Barren Land (Rock/Sand/Clay)
41 Deciduous Forest
42 Evergreen Forest
43 Mixed Forest
52 Shrub/Scrub (area dominated by shrubs)
71
Grassland/Herbaceous (unmaintained, dominayed by
herbaceous vegetation
81 Pasture/Hay (land intended for livestock grazing)
82 Cultivated Crops (annual crops/orchards/vineyards)
90
Woody Wetlands (areas where forest or shrubland
vegetation accounts for greater than 20% of vegetative cover
and the soil or substrate is periodically saturated with or
covered with water)
91
Emergent Herbaceous Wetlands (areas where perennial
herbaceous vegetation accounts for greater than 80% of
vegetative cover and the soil or substrate is periodically
saturated with or covered with water)
29

The NCLD 2001 data was processed as the NLCD 2011 data with the city boundaries and
denoting the pertaining classes. The Landsat 7 ETM+ imagery needed to be composited for a
natural color image. This was completed in ArcMap with the Composite tool using Bands 5, 4,
and 3. Now having obtained a natural color image of this section of Southern California, the
same comparison was made for image quality. A boundary shapefile was utilized to narrow in
on the two study areas. The clip tool was used to focus only on the two cities using the boundary
shapefile as an input feature on the NLCD imagery. With this achievement, a comparison can
now be done with both 2011 and ultimately, 1992.
The same process was completed for the data for 1992. The only difference here is that
Landsat 5 MSS imagery was used to complete the natural color composite image instead of the
Landsat 7 ETM+ since the Landsat 7 satellite was not in operation in 1992. Bands 4, 3, and 2
were used in this case to create the composite image.
The percent of change have been obtained and the results is discussed in the following
chapter with a comparison to the population change to understand the increase in residential and
light commercial land cover and decrease in heavy commercial and industrial land cover.

30

CHAPTER 4: RESULTS
The most land cover change in both cities was in the classes referring to developed urban land
cover, with little change occurring in the classes referring to natural features such as forests and
bodies of water. The classes Developed Open Space and Low Intensity generally showed a
decrease in percentage, which was reflected in the growth of the Developed Medium and High
Intensity classes. Also, areas classified as in the Shrub/Scrub class declined in size in both
Pasadena and Inglewood. Inglewood had few classes for comparison due to its urban landscape;
meanwhile, Pasadena had a few more classes for a more exciting examination.
The main error in the process was that when clipping the NLCD data to the city boundary
lines, the data was not contained to just the city boundaries. ArcMap also took into result the
pixels that made a square clip that included the entire city boundary. This means, in large part to
Pasadena with its irregular size, more data from outside the actually boundaries were taken into
effect.
Each of the following figures displays the mapped results for each city for the appropriate
year with a legend showing the classes, along with a pie chart that displays the percentage of
each class with a listing of the number of times a class is represented on the map in order of most
to least (not all classes are listed but all are displayed in the chart).
4.1 Comparison of Land Cover, 1992 to 2001
Since the classes did not match up perfectly, the following results show the best class
comparison between 1992 and 2001 with the classes of the Anderson classification method for
1992 to 2001. The classes are broader, therefore, showing less detail than the decision-tree
classification used to determine the land cover change from 2001 to 2011. These results still
31

prove to be valid as classifications grouped as a whole, such as urban development, natural
vegetation, and crop cover still gave a land cover change percentage.
4.1.1 1992 Classifications
The percent of land cover by class for 1992 can be seen in Table 7 and Table 8 for Pasadena and
Inglewood, respectably. The first column in each table are the Anderson I classes, the second
column is the number of pixels of each class resides in each clip for each city, and the third
column is the percentage of the total land cover from the amount of total pixels. The percentage
is what will be put into the SI calculation to reach the land cover percent change.

Table 7. Percentage of land cover by Anderson I class for Pasadena in 1992
Pasadena
Classifications
1992
Pixel
Count
Percent
of Land
Cover
Open Water 52 0.02%
Urban 105092 35.54%
Barren Land 1697 0.57%
Forest 59379 20.08%
Grassland/Shrub 129296 43.72%
Agriculture 186 0.06%
Wetlands 27 0.01%
Source: Derived from the NCLD data
Table 8. Percentage of land cover by Anderson I class for Inglewood in 1992
Inglewood
Classifications
1992
Pixel
Count
Percent
of Land
Cover
Open Water 2 0.00%
Urban 51955 88.52%
Barren Land 935 1.59%
Forest 1034 1.76%
Grassland/Shrub 4673 7.97%
Agriculture 96 0.16%
Wetlands 1 0.00%
Source: Derived from the NLCD data
32

Figure 4. Anderson I classification of land cover for Inglewood in 1992
33

Figure 5. Anderson I classification of land cover for Pasadena in 1992
34

In Inglewood, 88.52% of the land cover was urban based. The remaining land cover is
comprised mostly of grassland/shrub, forest, barren land, and with agriculture representing a
mere 0.16%. As a whole in Pasadena, the urban land cover was 35.54% of the total land cover.
This also results in a 64.46% of non-urban cover in the city. That means most of the city was
still covered in forest, shrub, and cropland. However, as stated above, much of that vegetation
data comes from outside the city boundary. Figure 4 displays the resulting clip for Inglewood
and for Pasadena in Figure 5.
Both maps are weighted in red which represents the impervious land cover, and orange to
represent the urban and recreational turf to make up the urban land cover classes.
Pasadena contains more classes over Inglewood because of Pasadena’s geographical location
next to the San Gabriel Mountains. Visually inside the boundary lines, thirteen natural land
cover classes are represented in Pasadena and only nine in Inglewood. The interesting thing is
that Inglewood does have more natural land cover classes than hypothesized. The classes
representing forest and agriculture were surprising results for the heavy urbanized city.
4.1.2 2001 Classification
Against the data from 1992, the classes for 2001 had to be presented in the same way, even with
the new land cover classification system that took place in 2001. Tables 9 and 10 display the
results of the 2001 Anderson I classification for each study area. Visually, the Agriculture class
is outside the city boundary, therefore, those results are dismissed from comparison, with its
0.17% not influencing the results greatly. Overall, Inglewood is covered in red or impervious
land cover. Pasadena, on the other hand, is geographically diverse and contains more natural
land cover classes.

35

Figure 6. Anderson I classification of land cover for Inglewood in 2001
36

Figure 7. Anderson I classification of land cover for Pasadena in 2001
37

Table 9. Percentage of land cover by Anderson I class for Pasadena in 2001
Pasadena
Classifications
2001
Pixel
Count
Percent
of Land
Cover
Open Water 48 0.02%
Urban 170210 57.59%
Barren 290 0.10%
Forest 28668 9.70%
Grassland/Shrub 95924 32.45%
Agriculture 0 0.00%
Wetlands 422 0.14%
Source: Derived from the NLCD data

Table 10. Percentage of land cover by Anderson I class for Inglewood in 2001
Inglewood
Classifications
2001
Pixel
Count
Percent
of Land
Cover
Open Water 10 0.02%
Urban 58049 98.90%
Barren 7 0.01%
Forest 0 0.00%
Grassland/Shrub 529 0.90%
Agriculture 101 0.17%
Wetlands 0 0.00%
Source: Derived from the NLCD data

The city is majority urban land cover at 57.59%, with the remaining totals being natural
land cover which are mostly tallied from the land cover outside the city boundary. The natural
land cover classes do exist inside the boundaries so those classes were not removed from the
comparison.
4.1.3 1992-2001 Land Cover Percent Changes
The difference in color representations on the maps from 1992 and 2001 displays the variations
of the two classification systems used. In Inglewood, there is a visual loss of natural classes
38

represented from the 1992 map to the 2001 map. There are also appears to be open water
represented on the Inglewood 2001 map that was not on the Inglewood 1992 map.
Table 11 shows the percent land cover change by classes for Pasadena and Table 12 for
Inglewood. The percent changes noted in red reflect a negative change from 1992 to 2001. The
total urban percent change for Inglewood was 10.38% and the percent change in Pasadena was
22.05%.

Table 11. Land cover percent change of Pasadena from 1992 to 2001. Red numbers
highlight land cover change that decreased from 1992 to 2001.
Pasadena
Classifications
1992-2001
Percent
1992
Percent
2001
Percent
Change
Open Water 0.02% 0.02% 0.00%
Urban 35.54% 57.59% 22.05%
Barren 0.57% 0.10% 0.47%
Forest 20.08% 9.70% 10.38%
Grassland/Shrub 43.72% 32.45% 11.27%
Agriculture 0.06% 0.00% 0.06%
Wetlands 0.01% 0.14% 0.13%

Table 12. Land cover percent change of Inglewood from 1992 to 2001. Red numbers
highlight land cover change that decreased from 1992 to 2001.
Inglewood
Classifications
1992-2001
Percent
1992
Percent
2001
Percent
Change
Open Water 0.00% 0.02% 0.02%
Urban 88.52% 98.90% 10.38%
Barren 1.59% 0.01% 1.58%
Forest 1.76% 0.00% 1.76%
Grassland/Shrub 7.97% 0.90% 7.07%
Agriculture 0.16% 0.17% 0.01%
Wetlands 0.00% 0.00% 0.00%

39

4.2 Comparison of Land Cover, 2001 to 2011
The results for this comparison are more accurate and detailed due to the use of the same
decision-tree classification system in both years that accompanied the NLCD data. Direct class
comparisons were completed because of this fact, rather than the broad urban land cover classes.
The land cover percentage results of the 2001 classes were previously explained in the former
subchapter, but will now go into more detail in this chapter for more accurate comparison with
the 2011 data.
4.2.1 2001 Classifications
The urban land cover classes for 2001 and 2011 are Developed Open Space (21), Developed
Low Intensity (22), Developed Medium Intensity (23), and Developed High Intensity (24). The
total for total urban land cover for Inglewood is 98.89%. Figure 8 displays the land cover map
for Inglewood with the decision-tree classification showing densely covered in the pink-to-red
symbology for the urban land cover.

Table 13. Percentage of land cover by decision-tree classification for Pasadena in 2001
Pasadena Classifications 2001
Pixel
Count
Percent
of Land
Cover
Open Water 48 0.02%
Developed Open Space 56905 19.25%
Developed Low Intensity 68920 23.32%
Developed Medium Intensity 40419 13.68%
Developed High Intensity 3966 1.34%
Barren Land 290 0.10%
Deciduous Forest 7 0.00%
Evergreen Forest 25479 8.62%
Mixed Forest 3182 1.08%
Shrub/Scrub 92711 31.36%
Grassland/Herbaceous 3213 1.09%
Woody Wetlands 422 0.14%
Source: Derived from the NLCD data
40

Figure 8. Decision-tree classification of land cover for Inglewood in 2001
41

Figure 9. Decision-tree classification of land cover for Pasadena in 2001
42

Table 14. Percentage of land cover by decision-tree classification for Inglewood in 2001
Inglewood Classifications 2001
Pixel
Count
Percent
of Land
Cover
Open Water 10 0.02%
Developed Open Space 2138 3.64%
Developed Low Intensity 6207 10.57%
Developed Medium Intensity 33767 57.54%
Developed High Intensity 15937 27.15%
Barren Land 7 0.01%
Grassland/Herbaceous 529 0.90%
Pasture/Hay 101 0.17%
Source: Derived from the NLCD data

The other classes represented in the Inglewood classification are Open Water (11), Barren Land
(31), Grassland/Herbaceous (71), and Pasture/Hay (81). Pasadena, represented in Figure 9,
shows more natural feature classes due to its geographically diverse location.
4.2.2 2011 Classifications
In 2011, Pasadena still had its geographically diverse land cover which gives it less of an urban
land cover percentage. As seen in Table 15, the city’s total urban percentage was 57.79%, with
most of the urban land cover having less than a 50% impervious surface. The percentages show
that Pasadena is still covered in large tracts of forest and shrubland. According to Table 16,
Inglewood on the other hand had a 99.48% urban land cover, with the rest made up of
Grassland/Herbaceous at .51% and Barren Land at .01%.

43

Table 15. Percentage of land cover by decision-tree classification for Pasadena in 2011
Pasadena Classifications
2011
Pixel
Count
Percent
of Land
Cover
Open Water 54 0.02%
Developed Open Space 55676 18.83%
Developed Low Intensity 67795 22.94%
Developed Medium Intensity 42397 14.34%
Developed High Intensity 4940 1.67%
Barren Land 234 0.08%
Deciduous Forest 7 0.00%
Evergreen Forest 25499 8.63%
Mixed Forest 3174 1.07%
Shrub/Scrub 92632 31.34%
Grassland/Herbaceous 2738 0.93%
Woody Wetlands 435 0.15%
Source: Derived from the NLCD data
Table 16. Percentage of land cover by decision-tree classification for Inglewood in 2011
Inglewood Classifications
2001
Pixel
Count
Percent
of Land
Cover
Open Water 0 0.00%
Developed Open Space 2072 3.53%
Developed Low Intensity 5841 9.95%
Developed Medium Intensity 34025 57.97%
Developed High Intensity 16451 28.03%
Barren Land 7 0.01%
Grassland/Herbaceous 300 0.51%
Pasture/Hay 0 0.00%
Source: Derived from the NLCD data

44

Figure 10. Decision-tree classification of land cover for Inglewood in 2011
45

Figure 11. Decision-tree classification of land cover for Pasadena in 2011
46

4.2.3 2001-2011 Land Cover Percent Changes

The results for the comparison of 2001 to 2011 followed the same hypothesis that the
urban land cover would change to higher developed intensities. Tables 17 and 18 shows the
percent changes for the classes of each city, with the red percent changes representing the
negative changes. Both cities have a negative percent change in the Developed Open Space and
Developed Low Intensity classes, which is made up for in the Developed Medium and High
Intensity classes. Inglewood’s highest percent change was in the Developed High Intensity class
and Pasadena’s highest percent change was in the Developed Medium Intensity class.

Table 17. Land cover percent change in Pasadena from 2001 to 2011. Red numbers
highlight land cover change that decreased from 2001 to 2011.
Pasadena Classifications
2001-2011
Percent
2001
Percent
2011
Percent
Change
Open Water 0.02% 0.02% 0.00%
Developed Open Space 19.25% 18.83% 0.42%
Developed Low Intensity 23.32% 22.94% 0.38%
Developed Medium Intensity 13.68% 14.34% 0.66%
Developed High Intensity 1.34% 1.67% 0.33%
Barren Land 0.10% 0.08% 0.02%
Deciduous Forest 0.00% 0.00% 0.00%
Evergreen Forest 8.62% 8.63% 0.01%
Mixed Forest 1.08% 1.07% 0.01%
Shrub/Scrub 31.36% 31.34% 0.02%
Grassland/Herbaceous 1.09% 0.93% 0.16%
Woody Wetlands 0.14% 0.15% 0.01%

Inglewood also had a negative percent change in the two classes that represent the natural
land cover. The Grassland/Herbaceous class dropped thirty-nine percent. Even though
previously stated that the Pasture/Hay class was visually outside the city boundary on the map, a
note was taken in the fact that the class has dropped to a zero percentage.
47

Table 18. Land cover percent change in Inglewood from 2001 to 2011. Red numbers
highlight land cover change that decreased from 2001 to 2011.
Inglewood Classifications
2001-2011
Percent
2001
Percent
2011
Percent
Change
Open Water 0.02% 0.00% 0.02%
Developed Open Space 3.64% 3.53% 0.11%
Developed Low Intensity 10.57% 9.95% 0.62%
Developed Medium Intensity 57.54% 57.97% 0.43%
Developed High Intensity 27.15% 28.03% 0.88%
Barren Land 0.01% 0.01% 0.00%
Grassland/Herbaceous 0.90% 0.51% 0.39%
Pasture/Hay 0.17% 0.00% 0.17%

Pasadena also saw a loss in natural land cover. There was a negative percent change for
the Barren Land class, Mixed Forest class, Shrub/Scrub class, and Grassland/Herbaceous class.
This negative percent change added to the growth of the urban land cover.
4.3 Pasadena vs. Inglewood
As stated before, these two cities are nothing alike other than the fact that they reside within the
Los Angeles metropolitan area. The graph seen in Figure 12 displays the land cover change of
the two cities, changing at different rates over the twenty year period. Pasadena is larger with a
landscape that contains both mountains with large tracts of natural forest and shrubland and
urban features, while Inglewood is a highly urbanized city with little vegetation, according to the
land cover data. Yet, both cities grew in urban land cover over the twenty year study period,
losing parts of their landscape that is classified as natural land cover.
The presumption of this study was to understand Inglewood as an always urban dominate
city. This is true since the early nineties; however, there was more land cover classified as non-
urban land cover from the 1992 data, which has been replaced with almost 100% urban
landscape according to the 2011 land cover data. The fact that there were non-urban
48

classifications in Inglewood was an eye-opener. The city has changed to highly urbanized rather
than always being highly urbanized in this time period. In Pasadena, the results occurred as
expected. The city’s already urbanized center just became heavier urbanized while keeping most
of it natural land cover intact. The cities did not gain much land area in urban land cover,
however, where there was existing urban land cover, it got more dense.

Figure 12. Bar graph displaying the land cover change by classification over the
twenty year period.

0.00%
10.00%
20.00%
30.00%
40.00%
50.00%
60.00%
70.00%
80.00%
90.00%
100.00%
Pasadena Inglewood Pasadena Inglewood Pasadena Inglewood
1992 2001 2011
Open Water
Urban
Barren Land
Forest
Grassland/Shrub
Agriculture
Wetlands
49

CHAPTER 5: DISCUSSION AND REVIEW
The urban land cover change for the study areas by analyzing NLCD data reflected the real
world change that satellite imagery can provide. Working with the challenge of comparing
different classification systems proved to be a challenge, but the NLCD data was able to be
roughly compared to understand the resulting land cover changes. The Urban Land Expansion
Index accurately compared the percentage or urban land change to demonstrate the increases and
decreases of certain classes.
5.1 Observations and Limitations
The overall comparison of the NLCD data showed the expected urban land cover increase in
both Pasadena and Inglewood, at different rates respectably. Pasadena increased urban land
cover in the lower developed intensities while Inglewood increased in the higher developed
intensities. This means that Pasadena grew more residentially, in both single and multi-family
structures, and in light commercial, along with urban and recreational turf. This is seen on the
maps in the growth of the lighter red colors. Pasadena maintains counts in the natural classes in
the more mountainous areas of the city where urban sprawl cannot spread into. However,
knowing that Pasadena is titles a “Tree City USA” as mentioned in Chapter One, there does not
seem to be green to represent trees or grass mixed in with the red urban areas. This is accounted
in the classification process. Each pixel is classified as the most determining characteristic and
prominent features. This means, for an example, a pixel with an office building with a couple of
trees around the outside would get classified as urban due to the office building being the
prominent feature of that pixel.
As for Inglewood, the already highly dense urban center grew in commercial and
industrial properties, which reflects in the maps as getting more and more dark red. There is less
50

than a one percent representation of non-urban land cover in this metropolitan city. The decrease
in the Developed Low Intensity class reflects the decrease of the population that occurred from
2000 to 2010. The city is being built up for more commercialism rather than housing.
The findings of this study are that both cities seem to have peaked with land cover
change, reflected in the minimal change from 2001 to 2011. The only noticeable changes are
from lower intensity developed urban classes into higher intensity developed classes. The city’s
urban planning policies have restrictions, keeping what is residential, or Developed Low
Intensity, from conversion into commercial, so there is no real land cover change. Only already
built-up areas becoming more dense, and that is what the policies are more concerned about.
There does not look like there will be much land cover change in the next decade for either city
if the past decade is any indication.
5.2 Significance
The results of the Urban Land Expansion Index analysis demonstrated that not all data are
created equal. The comparison of the 1992 NLCD data to that of 2001 was a challenge. The
information on the NLCD data website warned against the direct comparison since the
classification systems were different. However, they did not say it was impossible. The method
of using the SI and grouping classes together worked to show land cover changes.
The urban land cover change for both Inglewood and Pasadena increased as a whole over
the twenty years. Inglewood developed into a highly dense urban city; meanwhile is Pasadena
majorly urbanized, but still has concentrations of natural land cover. Pasadena has mountains on
its northern and western borders where the natural landscape is hard to urbanize so it is left in the
natural state and will most likely remain that way. In the case for Inglewood, the very little
51

barren land that remains will most likely be reclassified into one of the developed intensity
classes for the next NLCD dataset.
The results of comparison between 1992 and 2001 are not perfectly accurate because of
the nature of rough class grouping; therefore, the work completed by the MRLC which have
already completed the comparison should be utilized for more accurate answers. This study did
not take a look into the report done by the MRLC because the methodology was being tested to
see if the results would still work.
The second case for error in this study was the clipping of the imagery with the city
boundary shapefiles. The manner in which ArcMap clipped the NLCD data did not accurately
follow the boundaries of the study areas and included more data that made the classification
counts higher than they should be. A technique that replicates using the shapefiles as cookie
cutters will benefit the study so that the results are confined to only to the study areas.
5.3 Future Research
This study’s method can be replicated for other cities to compare themselves to either Pasadena
or Inglewood or a new city or cities to understand whether there is a need to update or develop
their urban planning. The case here is that the land cover data along with population statistics can
show the preliminary future, with the understanding that there are fluxes and those fluxes should
be accounted for also. An application of the socio-economic information to understand the urban
land cover changes could prove useful with understanding the reasons for urban land cover
growth. The land use policies created and used by both cities can benefit from understanding
the land cover changes that occurred to either changed or strengthen those policies that are in
place.
52

In a study completed by Robert W. Hoyer and Heejun Chang, they focused on the ever-
changing surface of the Earth in the hands of humankind to understand what has happened to
understand potential solutions for the future. The answer to their research question became
based on future population totals; more people living in the Oregon metropolitan area require
more living space. This leads to the involvement of research of population data to understand
what the population will amount to in study areas to help understand the growth of urban land
cover. Using U.S. Census data could help predict how Pasadena and Inglewood grow, or
decline, in population to better understand their future land cover change.
This process can be replicated, with errors in mind, in many cities and regions across the
United States because the NLCD data covers the entire country so the process can be easily
transferred to another study area. The process does not only work for urban land cover change,
rural areas can also benefit from this process to maintain cropland, forested areas, or wetlands
since the NLCD data has classifications for natural features. Tree cover data can be included to
help with finer detail of vegetation cover. Needless to say, this process can be easily replicated
for many situations with the concurrent use of NLCD data.

53

REFERENCES
ArcGIS, 10.1. ESRI. Redlands, CA.

Aspinall, Richard. 2004. Modelling land use change with generalized linear models: a multi-
model analysis of change between 1860 and 2000 in Gallatin Valley, Montana. Journal
of Environmental Management. No. 72 (2004): 91-103.

Bailey, Timothy J. and James B.M. Schick. Historical GIS: Enabling the collision of history
and geography. Social Science Review. No. 27 (August 2009): 291-296.

Berling-Wolff, Sheryl and Jianguo Wu. Modelling urban landscape dynamics: A case study
in Phoenix, USA. Urban Ecosystems. 17(2004): 215-240.

Bolstad, Paul. 2012. GIS Fundamentals: A First Text on Geographical Information Systems.
4
th
Ed. White Ebar Lake, MN: Eider Press.

Campbell, James B. and Randolph H Wynne. 2011. Introduction to Remote Sensing. 5
th
Ed.
New York: Guilford Press.

City of Inglewood. City History. http://www.cityofinglewood.org/about/city_history.asp
(accessed February 2015).

City of Inglewood. General Plan. 2014.

City of Pasadena. Land Use and Mobility Element Policies. 2014.

Cohen, Joel E. 1995. Population growth and earth’s human carrying capacity. Science
No. 269 (July 21, 1995): 341–346.

Cousins, Sarah A.O. 2001. Analysisof land-cover transitions based on 17
th
and 18
th
century
cadastral maps and aerial photographs. Landscape Ecology. No. 16 (2001): 41-54.

Deng, Jin S., Ke Wang, Yang Hong, and Jia G. Qi. 2009. Spatio-temporal dynamics and
evolution of land use change and landscape pattern in response to rapid urbanization.
Landscape and Urban Planning. No. 92 (2009): 187-198.

Foody, Giles M. 2002. Status of land cover classification accuracy assessment. Remote
Sensing of Environment. No. 80 (April 2002): 185-201.

Foresman, Timothy W., Pickett, Steward T.A., and Zipperer, Wayne C. 1997. Methods for
spatial and temporal land use and land cover assessment for urban ecosystems and
applications in the Greater Baltimore-Chesapeake Region. Ecosystems. No. 1 (1997):
201-216.

54

Fry, J.A., M.J. Coan, C.G. Homer, D.K. Meyer, and J.D. Wickman, 2009. Completion of the
National Land Cover Database (NLCD) 1992-2001 Land Cover Change Retrofit Product.
Open-file Report 2008-1379.

Gennaretti, Fabio, Maria Nicolina Ripa, Federica Gobattoni, Lorenzo Boccia, and Raffaele
Pelorosso. 2011. A methodology proposal for land cover change analysis using historical
aerial photos. Journal of Geigraphy and Regional Planning. No. 4 (September 2011):
542-556.

Homer, C. C., L. Yang Huang, B. Wylie and M. Coan. 2004. Development of a 2001
National Landcover Database for the United States. Photogrammetric Engineering and
Remote Sensing, No. 70 (July 2004): 829-840.

Hoyer, Robert W. and Heejun Chang. 2014. Development of Future Land Cover Change
Scenarios in the Metrpolitan Fringe, Oregon, U.S., with Stakeholder Involvement. Land,
No.3 (March 2014):322-341.

Jenerette, G. Daniel and Jianguo Wu. 2001. Analysis and simulation of land-use change in
the central Arizona-Phoenix region, USA. Landscape Ecology. No.16 (2001): 611-626.

Jin, S., L. Yang, P. Danielson, C. Homer J. Fry, and G. Xian. 2013. A comprehensive change
detection method for updating the National Land Cover Database to circa 2011. Remote
Sensing of Environment, No. 132 (2013): 159 - 175.

Los Angeles Almanac. General Population by City, Los Angeles County.
http://www.laalmanac.com (accessed April 7, 2014).

Luck, Matthew, and Jianguo Wu. 2002. A gradient analysis of urban landscape pattern: a case
study from the Phoenix Metropolitan region, Arizona, USA. Landscape Ecology. No. 17
(2002): 327-339.

Meiyappan, Prasanth and Atul K. Jain. 2012. Three distinct global estimates of historical
land-cover change and land-use conversions for over 200 years. Frontiers of Earth
Science, No. 6 (July 2012): 122-139.

National Space and Aeronautics Administration. Technical Details.
http://landsat.gsfc.nasa.gov (accessed May 2015).

Pickett, Steward. T. A., William R. Jr. Burch, Shawn E. Dalton, Timothy W. Foreman,
J. Morgan Grove, and Rowan Rowntree. 1997. A conceptual framework for the study of
human ecosystems in urban areas. Urban Ecosystems. No. 1 (1997): 185–199.

Riebsame, W. E., W.J. Parton, K.A. Galvin, I.C. Burke, L. Bohren, R. Young, and E.0 Knop
1994. Integrated modeling of land use and cover change. Bioscience. No. 44 (May 1994):
350–356.

55

Shenghe, Liu, Sylvia Prieler, and Li Xiubin. 2002. Spatial patterns of urban land use growth
in Beijing. Journal of Geographical Sciences. No. 12 (2002): 266-274.

Simpson, James R. 1993. Urbanization, agro-ecological zones and food production
sustainability. Outlook on Agriculture. No. 22 (1993): 233–239.

Tian, Guangjin, Jiyuan Liu, Yichun Xie, Zhifeng Yang, Dafang Zhuang, and Zheng Niu.
2005. Analysis of spatio-temporal dynamic pattern and driving forces of urban land in
China in 1990s using TM images. Cities. No. 22 (2005): 400-410.

U.S. Census Bureau: State and County QuickFacts. 2015. Data derived from Population
Estimates, American Community Survey, Census of Population and Housing, County
Business Patterns, Economic Census, Survey of Business Owners, Building Permits,
Census of Governments. http://quickfacts.census.gov/qfd/states/06/0656000.html
(accessed February 2015).

Veldkamp, A. and Lambin, E.F. 2001. Predicting land-use change. Agriculture, Ecosystems,
and Environment. No. 85 (2001): 1-6.

Vitousek, P. M. 2004. Beyond global warming: Ecology and global change. Ecology. No. 75
(October 1994): 1861–1876.

Vogelmann, James E., Stephen M. Howard, Limin Yang, Charles R. Larson, Bruce H. Wylie,
and Nick Van Driel. 2001. Completion of the 1990s National Land Cover Data Set for
the Conterminous United States from Landsat Thematic Mapper Data and Ancillary Data
Sources. Photogrammetric Engineering & Remote Sensing. No. 67 (June 2001): 650-662.

Wu, Jianguo and John L. David. 2002. A spatially explicit hierarchical approach to modeling
complex ecological systems: theory ad applications. Ecological Modeling. No. 153
(2002): 7-26.

Zipperer, Wayne C., Jianguo Wu, Richard V. Pouyat, and Steward T.A. Pickett. 2000. The
application of ecological principles to urban and urbanizing landscapes”. Ecological
Applications. No. 10 (2000): 685-688.

Abstract (if available)

Abstract Imagery and spatial data collected from different tools and satellite technologies have been used to complete land cover change studies at the scale of cities, countries and continents. Different methodologies have been used to complete these studies, dependent upon the technology and information available to complete land cover change. In this thesis, urban land cover has been analyzed by applying Landsat satellite imagery to spatial analysis as a way to examine land cover changes in Pasadena, California and Inglewood, California from 1992 to 2011. The objective for this study has been to review spatial data collected from Landsat data in order to understand urban land cover change in each city. Spatial data collected from the National Land Cover Database (NLCD) have been pre‐processed with color infrared composite creation and image classification tools to show land cover. Imagery from Landsat 4 has been used to help compare land cover change from 1992 to 2001 since classifications of the NLCD were different in both years. The resulting maps display the land cover changes over time from the effective application of imagery analysis to complete a pattern of land cover change over the time of twenty years. The study’s findings demonstrate that cities in the same metropolitan center can have similar urban growth patterns even when they have geographically diverse landscapes. These findings underscore the importance of understanding urban grown patterns when planning for urban.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Using census data, urban-land cover classification, and dasymetric mapping to measure urban growth of the lower Rio Grande Valley, Texas

PDF

Quantifying changes in monarch butterfly habitat in California

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

Comparing Landsat7 ETM+ and NAIP imagery for precision agriculture application in small scale farming: a case study in the south eastern part of Pittsylvania County, VA

PDF

Using Landsat and a Bayesian hard classifier to study forest change in the Salmon Creek Watershed area from 1972–2013

PDF

Does the Bay Area have a social center? Delimiting the postmodern urban center of the San Francisco Bay Area

PDF

Investigation and analysis of land use/land cover to determine the impact of policy developments in cities

PDF

Use of remotely sensed imagery to map sudden oak death (Phytophthora ramorum) in the Santa Cruz Mountains

PDF

Social media to locate urban displacement: assessing the risk of displacement using volunteered geographic information in the city of Los Angeles

PDF

Effects of building modifications and municipal policies on green cover in Los Angeles County

PDF

Redefining urban food systems to identify optimal rooftop community garden locations: a site suitability analysis in Seattle, Washington

PDF

Developing a replicable approach for the creation of urban climatic maps for urban heat island analysis: a case study for the city of Los Angeles, California

PDF

Distribution and correlates of feral cat trapping permits in Los Angeles, California

PDF

A comparison of pixel-based, object-based, deep learning, and data fusion image classification approaches in delineating urban tree canopy in City Terrace

PDF

Effect of spatial patterns on sampling design performance in a vegetation map accuracy assessment

PDF

Historical temperature trends in Los Angeles County, California

PDF

Disparities in food access: an empirical analysis of neighborhoods in the Atlanta metropolitan statistical area

PDF

Questioning the cause of calamity: using remotely sensed data to assess successive fire events

PDF

Integrating Landsat and California pesticide exposure estimation at aggregated analysis scales: accuracy assessment of rurality

PDF

Analysis of future land use conflict with volcanic hazard zones: Mount Rainier, Washington

Asset Metadata

Creator Crowther, Richard J., Jr. (author)

Core Title A comparison of urban land cover change: a study of Pasadena and Inglewood, California, 1992‐2011

School College of Letters, Arts and Sciences

Degree Master of Science

Degree Program Geographic Information Science and Technology

Publication Date 07/08/2015

Defense Date 05/11/2015

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

Tag classification,GIS,Inglewood,land cover,land cover change,landsat,NLCD,OAI-PMH Harvest,Pasadena,spatial,urban land cover

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Warshawsky, Daniel N. (committee chair), Lee, Su Jin (committee member), Ruddell, Darren M. (committee member)

Creator Email rcrowthe@usc.edu,rcrowtherjr@gmail.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c3-587225

Unique identifier UC11299596

Identifier etd-CrowtherRi-3561.pdf (filename),usctheses-c3-587225 (legacy record id)

Legacy Identifier etd-CrowtherRi-3561.pdf

Dmrecord 587225

Document Type Thesis

Format application/pdf (imt)

Rights Crowther, Richard J., Jr.

Type texts

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

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

Repository Name University of Southern California Digital Library

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