University of Southern California Dissertations and Theses

The impact of severe coastal flooding on economic recovery disparities: a study of New Jersey communities following Hurricane Sandy

(USC Thesis Other)

The impact of severe coastal flooding on economic recovery disparities: a study of New Jersey communities following Hurricane Sandy

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

The Impact of Severe Coastal Flooding on Economic Recovery Disparities:
A Study of New Jersey Communities Following Hurricane Sandy

by

Megan Rae Kelly

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)

May 2022

Copyright © 2022 Megan Rae Kelly

ii

To Sam and my loved ones
iii
Acknowledgments
I am grateful to my mentor, Dr. Ruddell, for his concise and pointed guidance and my other
committee members, Dr. Vos and Dr. Fleming, for their expertise and inspiration. Additionally, I
appreciate my employer, Tetra Tech, funding part of my education and understanding the careful
balance necessary to succeed.

iv
Contents
Dedication ....................................................................................................................................... ii
Acknowledgments.......................................................................................................................... iii
List of Tables ................................................................................................................................. vi
List of Figures ............................................................................................................................... vii
Abbreviations ................................................................................................................................. ix
Abstract .......................................................................................................................................... xi
Chapter 1 Introduction .................................................................................................................... 1
1.1. New Jersey and Flooding ....................................................................................................2
1.2. Motivation ...........................................................................................................................5
1.2.1. Global Climate Change and SLR ...............................................................................6
1.2.2. Maintaining Wealth Distribution in Coastal Communities .......................................7
1.3. Research Goals..................................................................................................................10
1.4. Study Organization ...........................................................................................................10
Chapter 2 Background .................................................................................................................. 11
2.1. Disaster Recovery Research .............................................................................................11
2.2. Modeling Flood Vulnerability ..........................................................................................13
2.3. Recovery Metrics and Spatial Statistics ............................................................................15
Chapter 3 Data and Methodology ................................................................................................. 19
3.1. Data ...................................................................................................................................19
3.1.1. Hurricane Sandy Surge Boundary ...........................................................................20
3.1.2. ACS 5-Year Estimates and Accuracy ......................................................................21
3.2. Research Design................................................................................................................23
3.2.1. Impact Zone Classification ......................................................................................23
3.2.2. Physical Evaluation ..................................................................................................24
v
3.2.3. Comparison of Years Before and After Disaster .....................................................25
Chapter 4 Results .......................................................................................................................... 31
4.1.1. Impact Zone Classification ......................................................................................31
4.1.2. Physical Evaluation ..................................................................................................32
4.1.3. Comparison of Years Before and After Disaster .....................................................35
Chapter 5 Discussion and Conclusion .......................................................................................... 58
5.1. Limitations and Future Research ......................................................................................58
5.2. Conclusions .......................................................................................................................62
References ..................................................................................................................................... 64
Appendix A ACS 5-Year Estimate Attributes .............................................................................. 67

vi
List of Tables
Table 1 Data dictionary of input spatial and tabular datasets for the study analysis. ................... 20
Table 2 Physical Evaluation of Elevation and Slope per Impact Zone. ........................................ 33
Table 3 Population Evaluation per Impact Zone for ACS Years 2006 – 2010 and 2014 – 2018. 35
Table 4 Median Household Income Evaluation per Impact Zone for ACS Years 2006 – 2010 and
2014 – 2018................................................................................................................................... 40
Table 5 Income Classes Evaluation per Impact Zone for ACS Years 2006 – 2010 and 2014 –
2018............................................................................................................................................... 43
Table 6 Housing Evaluation per Impact Zone for ACS Years 2006 – 2010 and 2014 – 2018. .... 46
Table 7 Spatial Autocorrelation (Global Moran's I) of Variable Difference from 2010-2018. .... 57

vii
List of Figures
Figure 1 Study Area of New Jersey, United States. ........................................................................ 2
Figure 2 New Jersey's Population 2000-2019. ................................................................................ 4
Figure 3 The Disaster Life Cycle (Flanagan et al. 2011) ................................................................ 8
Figure 4 New Jersey's Hurricane Sandy Surge Boundary. ........................................................... 21
Figure 5 Surface Elevation of New Jersey. ................................................................................... 25
Figure 6 Hurricane Sandy Surge Impact Zones of New Jersey. ................................................... 32
Figure 7 Average Mean Elevation per Census Tract in New Jersey. ........................................... 34
Figure 8 Average Mean Slope per Census Tract in New Jersey................................................... 34
Figure 9 Population Change Graph, 2006 – 2010 and 2014 – 2018. ............................................ 36
Figure 10 2010 Population Density per Census Tract in Hurricane Sandy Surge Boundary. ...... 37
Figure 11 Average Median Household Income Graph, 2006-2010 and 2014-2018. .................... 41
Figure 12 Lowest Median Household Income Graph, 2006-2010 and 2014-2018. ..................... 41
Figure 13 Highest Median Household Income Graph, 2006-2010 and 2014-2018. ..................... 41
Figure 14 Persons Below Poverty Percentage Graph, 2006-2010 and 2014-2018. ...................... 41
Figure 15 Lower-Income Households Percentage Graph, 2006 – 2010 and 2014 – 2018. .......... 44
Figure 16 Middle-Income Households Percentage Graph, 2006 – 2010 and 2014 – 2018. ......... 44
Figure 17 Upper-Income Households Percentage Graph, 2006 – 2010 and 2014 – 2018. ........... 44
Figure 18 Percentage of Owner-Occupied Housing Units Graph, 2006-2010 and 2014-2018. ... 47
Figure 19 Percentage of Renter-Occupied Housing Units Graph, 2006-2010 and 2014-2018. ... 47
Figure 20 Percentage of Vacant Housing Units Graph, 2006-2010 and 2014-2018. ................... 47
Figure 21 Hot Spot Analysis of Population Difference from 2010 – 2018. ................................. 49
Figure 22 Hot Spot Analysis of Population Density Difference from 2010 – 2018. .................... 49
Figure 23 Hot Spot Analysis of Median Household Income Difference from 2010 – 2018. ....... 51
Figure 24 Hot Spot Analysis of Persons Below Poverty Difference from 2010 – 2018. ............. 51
viii
Figure 25 Hot Spot Analysis of Lower-Income Household Difference from 2010 - 2018. ......... 52
Figure 26 Hot Spot Analysis of Middle-Income Household Difference from 2010 - 2018. ........ 53
Figure 27 Hot Spot Analysis of Upper-Income Household Difference from 2010 - 2018. .......... 53
Figure 28 Hot Spot Analysis of Owner-Occupied Housing Difference from 2010 – 2018. ........ 55
Figure 29 Hot Spot Analysis of Renter-Occupied Housing Difference from 2010 – 2018. ......... 55
Figure 30 Hot Spot Analysis of Vacant Housing Unit Difference from 2010 – 2018.................. 56

ix
Abbreviations
ACS American Community Survey
CV Coefficient of Variation
DEM Digital Elevation Model
FEMA Federal Emergency Management Agency
GIS Geographic information system
GISci Geographic information science
GPS Global Positioning System
HMA Hazard Mitigation Assistance
IPCC Intergovernmental Panel on Climate Change
LiDAR Light Detection and Ranging
MOE Margin of Error
MIZ Minor Impact Zone
MOTF Modeling Task Force
NASA The National Aeronautics and Space Administration
NDRF National Disaster Recovery Framework
NHRAP Natural Hazards Risk Assessment Program
NJ New Jersey
NJDEP New Jersey Department of Environmental Protection
NIZ None Impact Zone
SLR Sea-level rise
SrIZ Serious Impact Zone
SvIZ Severe Impact Zone
x
SVI Social Vulnerability Index
SSI Spatial Sciences Institute
USC University of Southern California
USGS The United States Geological Survey
xi
Abstract
Recent severe flooding caused by storms, such as Hurricane Sandy in 2012, has damaged
vulnerable coastal communities across the United States at an increasing occurrence and
severity. Not only do floods threaten lives and property, but they also alter the shape of a
community through imbalanced recovery among socially and economically vulnerable
populations. This concern begs the research question: what, if any, are the differences in
recovery between communities of different economic standing concerning flood inundation
levels after a severe coastal flooding event? Economic recovery disparity was investigated by
analyzing New Jersey's socio-economic structure before and after Hurricane Sandy according to
inundation depths categorized as impact zones: None (NIZ), Minor (MIZ), Serious (SrIZ), and
Severe (SvIZ). The research design was developed to (1) examine the physical exposure of
Hurricane Sandy across New Jersey; (2) investigate the socio-economic characteristics of New
Jersey communities before and after Hurricane Sandy; and (3) determine whether, or not,
proximity to severe flooding resulted in notable changes to citizen’s economic standing. The
analysis compared tabular data from 2010 and 2018 American Community Survey (ACS) 5-Year
Estimates using three evaluations: population, income, and housing. Results displayed variable
levels of impact throughout the entire study area from 2010 to 2018 regarding population,
income, and housing; however, results did not show statistically significant relationships
between economic recovery and flood inundation levels.
1

Chapter 1 Introduction
Affluent areas of a community historically recover more quickly from a natural disaster, while
their less privileged neighbors often struggle to regain their previous lifestyle. Recovery of a
community pivots on the residents’ ability to return to the previously affected area without
financial impediment (Howell and Elliott 2018). After a disaster, wealthier citizens can withstand
the reconstruction costs and increased cost of insurance compared to others forced to relocate.
The displacement of residents, particularly those who are most vulnerable, can transform the
structure of a community and increase existing inequalities (van Holm 2019). Maintaining a
socially and economically diverse community enables innovation and productive opportunities
for all members (CityObservatory 2018).
Although New Jersey is socio-economically diverse overall, income inequality is distinct
in pockets, particularly along the coast. From nuisance to severe flooding, the residents in these
communities have experienced increased instability due to the ongoing devastation to one of
their most important investments—their home. The growing presence of tropical storms and
rising sea levels in New Jersey intensifies the need to protect those most at risk of adverse
change in exposed coastal regions (Stocker et al. 2013). The most destructive natural disaster in
New Jersey’s history was Hurricane Sandy in 2012, where flood depths reached about 19 feet.
This study reviewed New Jersey's recovery after Hurricane Sandy to better understand the
relationship between a community’s economic recovery and physical exposure from a severe
flood event. Flood exposure and inundation depths from observed Hurricane Sandy storm surge
levels determined impact zones used to track notable changes before and after the storm.
Comparing 2010 and 2018 data helped illustrate recovery differences throughout New Jersey and
investigate the connection between recovery and vulnerability.
2

1.1. New Jersey and Flooding
In the Northeastern region of the United States, New Jersey is surrounded by New York,
Pennsylvania, and the Atlantic Ocean. The state's northwest region intersects with the foothills of
the Appalachian Mountains, providing a dramatic landscape change to the otherwise flat
topography (Figure 1). Surrounded by water bodies on three sides, New Jersey is a coastal state
with the Atlantic Ocean to the east, the Delaware Bay to the south, and the Delaware River to the
West. The flat topography and proximity to water bodies increase the state's risk of encountering
flood-related issues.

Figure 1. Study Area of New Jersey, United States.
3

Another factor that increases flooding is the impervious surface area due to the increased
amount of real estate along the coastline. Although barrier islands do not exhibit the highest
population density of the state, they are neighborhoods of dense single-family properties.
Nuisance flooding, or tidal flooding, is expected on barrier islands due to high levels of
impervious surfaces and high tide in low elevation areas. Naturally intended to protect the
mainland shores with dunes, the current barrier island complex in New Jersey is not sustainable
for annual hurricane seasons. Like Hoboken and Jersey City, other coastal cities flood due to
heavy precipitation and runoff that intensify with high tides and storm surges
(Athanasopoulou 2017).
In addition, New Jersey is home to two large metropolitan areas along the rivers: New
York City and Philadelphia. The bordering cities of New York City and Philadelphia also
influence New Jersey’s economy and population, with population tending to be the densest in
these regions (Figure 1). In 2020, New Jersey was deemed the most densely populated state in
the United States, with about 1,263 people per square mile (US Census n.d.). As part of the Tri-
State area, New Jersey has experienced a consistent rise in population due to public transport
accessibility for commuters working in the nearby major cities (Figure 2). Studies show that
New Jersey's exposure and population size, compared to other states, place them at an above-
average threat of coastal flooding, right behind Florida and Louisiana. As of 2000, New Jersey
had about 4% of the population, more than 350,000 people, at risk of a 100-year coastal flood,
making them the fourth most vulnerable state to coastal flooding (States at Risk 2015).
According to First Street Foundation (2020), almost 400,000 properties in New Jersey are at risk
of substantial flooding, which is determined as the inundation of one centimeter or more to a
structure in the 100-year storm zone and rounded to the nearest 100.
4

Figure 2. New Jersey's Population 2000-2019.
Compared to Florida and Louisiana, New Jersey had little experience with profound
coastal damage until October 29, 2012, when Hurricane Sandy landed on Brigantine, New
Jersey. In New Jersey alone, Hurricane Sandy damaged 346,000 homes, submerged 1,400 sea
vessels, affected 70 drinking water systems, impacted 80 wastewater treatment plants, and
eroded 194 miles of coastline (NJDEP 2015). The catastrophic damage to the once vibrant Jersey
Shore was a realization among lawmakers and planners to better prepare for future storms and
more frequent nuisance flooding (Bryner, Garcia-Lozano, and Bruch 2017). Observed water
levels were highest along the northern section of the Jersey Shore, from Long Branch to Toms
River (Figure 1). New Jersey’s barrier islands were inundated entirely or breached in some areas
due to storm surge and large waves (NJDEP 2015). The International Displacement Monitoring
Centre estimated that Hurricane Sandy displaced about 53,500 people three years after the
disaster (Bryner, Garcia-Lozano, and Bruch 2017).
Hurricanes are less impactful in the northeast than in the southern region of the United
States due to the storm losing strength while traveling away from its source. In New Jersey, most
5

hurricane damage derives from flood inundation rather than other aspects such as wind, heavy
rain, and storm surge; therefore, this study only examined flood inundation caused by Hurricane
Sandy. This project aims to review flood exposure and inundation depths from Hurricane Sandy
storm surge inundation levels to spot communities' most vulnerable socio-economic change
concerning coastal residents.
1.2. Motivation
According to the Fifth Assessment Report of the Intergovernmental Panel on Climate
Change (IPCC), global climate change severely affects tropical storm activity by rising sea
levels, increasing hurricane rainfall, and amplifying intensity. The most vulnerable locations of
coastal communities are at an increased risk due to sea-level rise (SLR) and coastal development
(Stocker et al. 2013). Prominent on the United States Eastern Coastline, SLR may exacerbate in
areas where land is "sinking" due to vertical motions of the Earth's crust. The region from North
Carolina to New Jersey is especially undergoing strong coastal subsidence from the last
deglaciation (Piecuch et al. 2018). Growing threats such as nuisance flooding and relative SLR
will be of more significant concern in this zone; therefore, mitigation will be essential (Jacobs et
al. 2018).
Post-disaster recovery is often the most telling way to analyze the severity of a storm.
Displacing a large portion of the population may destroy a community’s socio-economic
structure. About 7% of the population was still displaced three years after Hurricane Sandy,
primarily due to economic hardships (Bryner, Garcia-Lozano, and Bruch 2017). Retroactively
identifying the areas at risk of displacement helps preserve or encourage economic diversity
before a major disaster occurs. Income equality and socio-economic diversity benefit the
population in need and the entire community. The current model of pushing people out of
6

neighborhoods with economic and educational opportunities will, in turn, hinder the overall
economic growth of an area (Howell and Elliott 2018). If this pattern continues, coastal
communities will slowly transform into second homes for the wealthy; simultaneously, the
citizens with less income will be hit harder by financial setbacks, thus increasing wealth
inequality (van Holm 2019).
1.2.1. Global Climate Change and SLR
The growing risk of flood damage to coastal communities is becoming more apparent the
more commonplace they become. Experts predict that global climate change will increase the
frequency and severity of future tropical storm events resulting in more property damage and
loss of life. The most vulnerable locations of coastal communities are at an increased risk due to
SLR and impervious surfaces caused by coastal development (Stocker et al. 2013). The National
Aeronautics and Space Administration (NASA) Administrator, Bill Nelson, stated that increased
flooding poses an increased danger to low-lying areas near sea level due to the compounding
factors of the Moon's gravitational pull, SLR, and climate change. Low elevation coastal
communities have already seen the risks of high tide flooding events and expect to face worse
conditions with rising sea levels, and lunar amplify tides in the mid-2030s (Rasmussen 2021).
With the rate of SLR, constant and severe flood events will become a significant issue in
already sensitive areas. Also, Earth's crust plays a role in SLR, and it may affect how rising tides
damage different coastal zones throughout the world—taking into consideration the relative SLR
on the US East Coast will change how to manage New Jersey's coastal communities (Piecuch et
al. 2018). A report conducted by Rutgers University for the New Jersey (NJ) Climate Adaptation
Alliance gave projected estimates of the growing threat of SLR for the state of New Jersey. The
7

report estimated that the likely range of SLR by the year 2030 will be 0.6 – 1.0 ft with a 67%
probability (Kopp et al. 2016).
Climate change plays a significant role in defining how to reduce vulnerability and
increase resiliency to disaster risk. Five points of sustainable planning and proper design
improvement and integration help create a broader and deeper understanding of resilience to
reduce disaster risk. Those five points on vulnerability and resilience being (1) the
acknowledgment of non-qualitative characteristics of how a community avoids, reacts, and
recovers from a disaster; (2) the change in perspective to view hazards as resources or
opportunities to emphasize resilience; (3) the use of both absolute and proportional metrics
providing different impact results; (4) the focus on contextuality or localization of the area
affected; and (5) the need to examine long-term progress of recovery (Kelman, Gaillard, and
Mercer 2015).
1.2.2. Maintaining Wealth Distribution in Coastal Communities
Socio-economically diverse neighborhoods stimulate the economy by providing a variety
of needs, perspectives, and career opportunities and offer more robust social networks. Also,
affordable housing in areas with quality resources can break the trend of intergenerational
poverty and increase the likelihood of children from low-come backgrounds earning higher
wages than their parents. A study of the nation's most diverse, mixed-income neighborhoods
found that the communities remain diverse once established (CityObservatory 2018).
A community's existing bonds and stability help unite its residents and become more
resilient when facing hardships. If neglected in the wake of a disaster, wealth gaps are
emphasized, increasing inequality in susceptible areas (Howell and Elliott 2018). The dramatic
increase of a community's socio-economic status, otherwise known as gentrification, is a
8

sensitive matter to control when an area is recovering from a natural disaster. A hypothesis
dubbed the "recovery machine" suggests that social status before a disaster predetermines access
to resources and recovery (van Holm 2019). By this standard, less wealthy communities will, by
nature, struggle in recovering. In theory, this will result in a shift where lower-income residents
will endure financial troubles and leave the community creating an opportunity for developers to
buy properties to sell to wealthier individuals. In the case of a disaster, high populations of low-
income residents may not have the option of relocating, so they will stay in pockets of increasing
poverty—lowering the community's socio-economic status overall. Vulnerable coastal regions,
such as the New Jersey shore, have experienced the growing threat of severe flooding projected
to affect less wealthy residents the most.
Identifying hazards and how they pose a threat to the communities is one of the many
responsibilities of local governments before a disaster strikes. In attempts to assist in a
community's recovery, the Federal Emergency Management Agency (FEMA) established four
valuable phases called the Disaster Life Cycle for a successful and sustainable emergency
management plan—mitigation, preparedness, response, and recovery (FEMA 2017) (Figure 3).
Mitigation and resilience are the most important to minimize the other phases of the four steps.

Figure 3. The Disaster Life Cycle (Flanagan et al. 2011)
9

Frequently lower-income areas tend to be the communities with less protection and
preparedness. A study conducted in 2012 compared the recovery experience of the residents in
Houston, Texas, for pre- and post-hurricane Ike preparedness and questioned whether citizens
with first-hand experience of natural disasters are more prepared than those merely educated
before a disaster. Their findings show that preparedness was not uniform across the population of
Houston, TX, before hurricane Ike and minority populations reported poorer access to
information on preparedness and evacuation (Chen, Banerjee, and Lui 2012). Due to the
devastation, a lack of mitigation and preparedness leads to fewer citizens returning to their
communities. Five common factors – habitability of homes, affordability of housing, financial
burdens, sense of place and identity, and slow restoration of public services and facilities – have
been shown to influence displaced individuals' decision to return home after the disasters
(Bryner, Garcia-Lozano, and Bruch 2017).
To support the recovery phase of a disaster cycle, FEMA employs the National Disaster
Recovery Framework (NDRF) to support disaster-impacted areas throughout restoration and
redevelopment by providing Hazard Mitigation Assistance (HMA) grants and otherwise. The
goal of the NDRF is to administer services for economic recovery, health, social services,
housing, and natural and cultural resources after the disaster (Department of Homeland Security
2016). Conversely, the wealth gap evaluation caused by natural hazard damage has attributed to
unequal distribution of government assistance (Howell and Elliott 2018). Grants and government
funding greatly assist in recovery; however, the budget may not cover all the residents’ losses.
Establishing an in-depth planning strategy helps protect the most vulnerable in a more dynamic
approach. The United Nations Development Programme created new ways to address coastal
resilience using innovative financing mechanisms that fund natural infrastructure projects (Deutz
10

2018). These mechanisms help build economic growth before a disaster and could be valuable to
fund projects for less privileged areas in coastal communities.
1.3. Research Goals
The overarching purpose of this study was to examine the socio-economic characteristics
of existing coastal communities by developing an evaluation to understand the relationship
between recovery and physical exposure from a severe flood event. Previous research shows that
natural disaster recovery often results in changes in the community's economic characteristics.
The three research goals of this thesis are to:
• examine the physical exposure of Hurricane Sandy across New Jersey;
• investigate the socio-economic characteristics of New Jersey communities before and
after Hurricane Sandy; and
• determine whether proximity to severe flooding resulted in notable changes in
population, income, or housing.
1.4. Study Organization
The remaining study contains four additional chapters. Chapter Two provides
background on previous studies related to disaster recovery and the importance of Geographic
Information Systems (GIS) modeling. Using some of the proven methods, Chapter Three details
the techniques employed for testing the relationship between recovery and severe flooding, with
results documented in Chapter Four. In conclusion, Chapter Five reviews the results and explores
future studies.
11

Chapter 2 Background
This section examines assessment practices used to measure the vulnerability and recovery of
communities’ disasters, along with the value of using flood modeling and spatial statistics.
Related literature about disaster recovery helps explain how the study expands upon existing
knowledge of the socio-economic transformation of a coastal community.
2.1. Disaster Recovery Research
Natural disasters do not target certain socio-economic groups; however, there is
disproportional recovery across more impoverished populations. Current literature and research
fail to answer why some communities recover rather than others (Yabe et al. 2020) and the
enduring socio-economic effect on a community. Instead, existing literature examines
measurable costs such as property loss and migration of a population. These methods help create
a connection between immediate cause and effect, yet past recovery studies minimally examined
the long-term transformation of a community.
Yabe et al. (2020) examined human displacement for various study areas to understand
when and why communities recover. They used mobile phone data to better understand a
population's short-term and long-term migration patterns after a storm, as shown in a study for
Bangladesh after a cyclone. The authors pursued the connection between the recovery of
infrastructure systems and population movement after a major disaster by examining mobility
trajectories across three countries' mobile phone Global Positioning System (GPS) datasets
before, during, and after five significant disasters. The macro-scale of this study took another
step to downscale the analysis to counties and cities after modeling displacement rates to
understand the relationship between distance and duration of displacement after the disasters.
Results showed that most residents returned soon after the disaster, while some remained
12

displaced for extended periods. Factors such as population, median income, housing damage
rates, and length of infrastructure recovery time contributed to heterogeneity in short-term and
long-term displacement rates.
Although Yabe et al. (2020) provided essential clarity to the subject of recovery, the
question still stands about who is disproportionately affected by disasters according to recovery
rates. They suggested using household-level surveys to understand the demographic of people
unable to return to their previous residence. Examining recovery on a large scale helps visualize
the universal issues, but it does generalize the problem by comparing events with similar effects.
Yabe et al.'s (2020) work did not consider or present the effects of various disasters and
populations' backgrounds. Instead, the study defined demographics according to a single disaster
and used this measure to understand the relationship between the length of displacement and
existing socio-economic factors.
Another approach taken by researchers is to track neighborhood transformation
influenced by government and philanthropic-funded assistance. Although this is not the focus of
the study at large, touching upon post-disaster programs highlights the issues with current forms
of recovery by showing they may not be as effective as previously thought. Previous literature
reviews the effect of programs on residents' unstable financial situation.
In reviewing current disaster relief programs, a 2019 study from the University of
Colorado outlined the effect of the hurricane-caused flooding to Houston, TX, that resulted in
about $125 billion in damages to private and public property. Results show that averages mask
an essential assortment of experiences after disasters, challenging existing narratives of federal
disaster programs' effectiveness. They suggest that the current method of providing Small
Business Administration disaster loans and Federal Emergency Management Agency (FEMA)
13

grants can cause an adverse reaction to the communities they seek to support. In the methods, the
researchers compared the credit outcomes of Houston residents according to the amount of
flooding per Census block to track declining debt after flooding (Billings, Gallagher, and
Ricketts 2019). Tracking credit rather than income can be an effective way to examine the
success of recovery due to the financial setbacks to an unprepared population resulting from a
major disaster. Calculating the bankruptcy rate can be adopted for determining the change of
wealth in other coastal communities.
An analysis of the post-disaster assistance program, New York Rising Buyout, evaluated
willing participants and their property within vulnerable zones after Hurricane Sandy to test the
effectiveness of the state-run home buyout program. Literature shows that buyout programs
impact relocated residents and those who live in and around buyout areas. A GIS-based overlay
analysis compared the vulnerability of households with and without the buyout program. Change
in vulnerability calculated a Social Vulnerability Index (SVI) and the exposure rate to flooding
when participants moved from affected to unaffected neighborhoods. The study computed
relocation trends from the 323 participating households. Results showed that most participants
stayed near their original address. (McGhee, Binder, and Albright 2020). Creating similar study
groups is an effective way to monitor and compare movement while disaster recovery is
happening in real-time.
2.2. Modeling Flood Vulnerability
The previous literature presents the importance of incorporating socio-economic factors
when reviewing a community's recovery. Socio-economic and demographic factors likely affect
a community's recovery after a disaster due to the inability to effectively protect themselves and
their property before the flooding and a lack of capital for rebuilding in the recovery phase.
14

Exploring the factors that make a community vulnerable, such as average income, housing equity
values, disabilities, and age groups, will give a more accurate distribution of an area to predict
recovery.
Understanding previous literature that recovery is related to a community's wellbeing,
existing social and economic risks define such an index. Lichter and Felsenstein (2014) assessed
the socio-economic consequences of extreme coastal flooding events for Tel Aviv and a
collection of different-sized coastal communities in Israel. The study areas were all parts of the
low-elevation coastal zone, defined as the entire area below 10-m elevation and hydrologically
connected to the sea. Evaluating average income, disabilities, age groups, and the number of
vehicles per household determined social vulnerability; correlating house values with flood plain
elevations and gradients defined economic exposure. Also, the Gini coefficient calculated the
preflood income distribution of the vulnerable areas. This article used strategies of comparing
communities' vulnerability with various flood level scenarios using SLR calculations (Lichter
and Felsenstein 2014). Although Lichter and Felsenstein successfully demonstrated the link
between social exposure and physical susceptibility due to flooding, the Israeli-based study
called for data specific to the region. When reviewing the socio-economic vulnerability in the
United States, race, ethnicity, and gender are more connected due to disparity trends shown
throughout history (Howell and Elliott 2018).
In a review of socio-economic vulnerably after severe flooding within the United States,
the states bordering the Gulf of Mexico provide a more in-depth perspective of the population in
this area. Due to the common occurrence of hurricanes in the Gulf of Mexico and the southern
United States, there have been more opportunities to research socio-economic vulnerabilities in
this region. An article published assessing the US Gulf used GIS methods to analyze coastal
15

communities' vulnerability to hurricanes and flooding along the coast to highlight growing risks
to the economic health of the people in these communities. Flood exposure and an SVI were
analyzed and depicted overall community vulnerability. Population density and thirty social
vulnerabilities categorized and calculated the vulnerability of these communities. The results
showed that hurricanes and flooding could execrate vulnerability, and every component of the
comprehensive index plays a vital role in exposure (Shao et al. 2020). The study presented here
will compile several previous methods to construct a site-specific socio-economic vulnerability
index. The process of obtaining social vulnerability data sources on a census tract level
calculated the socio-economic risk of New Jersey's coastal communities. Applying the SVI
equations for a simplified matrix of factors census tract level will be helpful when defining
regions of coastal communities to compare later.
Presenting a community’s vulnerability using spatial thinking enhances the quality and
scope of a study by providing a deeper level of understanding to the data. A study by the
Universitas Negeri Malang assessed students’ knowledge of Indonesia’s and Iran’s disaster-
influenced impacts through GIS. They found that spatial patterns, linkage, and relationships
provide more decisive conclusions when analyzing vulnerability (Wahyuningtyas, Febrianti, and
Andini 2020). Visualizing the relationship between vulnerability and physical flood exposure in
New Jersey's coastal communities helps explain data otherwise absent from tabular data.
2.3. Recovery Metrics and Spatial Statistics
The recovery rate reflects how resilient a community is in the wake of a natural disaster.
Many articles look to relocation data to define recovery; however, there are various ways to
collect this type of data, e.g., postal service active delivery or population estimates. Additionally,
16

the community's location increases its inherent recovery shortcomings, which calculates socio-
economic and physical vulnerability.
Differences in a community’s vulnerability directly relate to an area's ability to recover.
A study of the 2004 tsunami in Thailand analyzed the relationship between factors of
vulnerability and developing adaptive strategies. The components of sensitivity and resilience in
testing exposure were used to test recovery. A community can achieve effective recovery with
proper adaptation strategies (Willroth et al. 2012). Due to the close relationship between
vulnerability and recovery, the study reapplied metrics used in exposure to the recovery metrics
in the review of New Jersey’s coastal communities.
Hurricane Katrina is arguably the most well-known flood disaster in US history. In a
2010 article by Finch, Emrich, and Cutter, they examined New Orleans' existing social
vulnerabilities and the level of flood exposure to understand inequities in recovery. Their study
used flood water levels of Hurricane Katrina to classify flood inundation at a census tract level,
then combined the results with social vulnerability. Dominant components of the SVI in this
study included race and class, young families, public housing, elderly, Hispanic immigrants,
special needs, and natural resources employment. Average flood depth recorded during the 2005
levee breach was assigned to each census tract, and level of damage was used to classify each
depth range as depth: None (0 ft), Low (<2 ft), Medium (2–4 ft), and High (>4 ft). The method of
determining residential relocation was to collect data from the US Postal Service of active
delivery locations and compare results before the hurricane in 2005 and after the hurricane in
2008. The study results showed that the SVI for New Orleans is significantly correlated with the
percentage of returned households by using Pearson's r as the correlation statistic (Finch, Emrich,
and Cutter 2010). Unlike New Jersey, New Orleans is recognized for its high risk of flooding and
17

socio-economic disparity compared to the nation. The methods of combining social vulnerability
and flood inundation shown in the study by Finch, Emrich, and Cutter give guidance on how to
spatially match regions in New Jersey's coastal communities if subregions are identified.
Two articles were examined to understand the practice of comparing similar regions and
how spatial matching with difference-in-differences measures could be used to understand the
recovery process in New Jersey's coastal communities. Holzer's 2017 study displayed the
effectiveness of the Minneapolis Neighborhood Revitalization Program by using census data for
neighborhood income, home value, rent, and vacancy rate. Neighborhood quality was analyzed
using difference-in-differences and hot spot analysis to compare similar neighborhoods. The core
of this methodology was that two control groups were designed to share standard propensity
scores. Minneapolis and St. Paul shared similar populations, racial and ethnic composition, size,
and urban form, vital components to his analysis (Holzer 2017). A similar article utilized the
spatial matching difference-in-differences estimator to test its effects on a 1998 flood event in
Laval, Québec, Canada. The method was used to isolate the impact of a change before and after
an exogenous difference between treatment and control areas (Dubé, AbdelHalim, and Devaux
2021). When matching regions in New Jersey's coastal communities, the control and treatment
groups examined previously will be determined by paring vulnerability rates and flood
inundation. Once the groups are defined, the relationship between vulnerability and relocation
will be examined, commonly done through statistical analysis.
By examining the relationship between vulnerability and rate of relocation after a
significant flood event, such as Hurricanes Katrina and Rita, researchers from Myers, Slack, and
Singelmann’s 2008 study gained an understanding of the community’s recovery in the US Gulf
region. They suggest that migration occurs because of social, economic, and geographic triggers.
18

The analysis used a dependent variable to calculate the percent of migration over one year after
the hurricanes by using county population percentage. A county-level SVI was calculated and
applied to macro levels of migration patterns. Another approach taken by researchers studying
recovery was to understand how location factors affect a community’s recovery. Spatial
statistics, such as regression analyses, show significant incidences of migration along the
hurricanes' path, and results showed disadvantaged populations were most likely to relocate
(Myers, Slack, and Singelmann 2008). Statistical analysis such as Pearson's r and regression
analyses help to assess social vulnerability, percent housing damage, and percent migration in
the New Jersey's study areas. To gain a complete understanding of socio-economic
characteristics and recovery, a study’s site-specific vulnerability factors must be evaluated, and
reliability tested.
19

Chapter 3 Data and Methodology
This study used a combination of tabular and spatial data to understand the correlation between
New Jersey communities' economic recovery and inundation levels after Hurricane Sandy's
coastal flooding event. Census tract polygons were consolidated to ensure only inhabited areas
were included in the study. A state-level classification of impact zones was linked to these areas
by estimating average inundation depth per census tract using the Sandy Surge boundary and
assigning zones from no impact to severe impact. Next, a physical evaluation was conducted to
incorporate elevation, slope, and affected census tracts to each impact zone. Finally, an analysis
of economic conditions before and after Hurricane Sandy in 2012 was performed using 2010 as
the baseline year and 2018 ACS 5-Year Estimates to calculate percent change.
Additionally, the reliability of the ACS was calculated using the coefficient of variation
(CV). In conclusion, spatial statistic tools were employed to understand the clustering patterns
and data relationships through Hot Spot Analysis. This chapter used the data and methodology to
define the inundation zones, data reliability, and data connections.
3.1. Data
The inputs for this analysis consisted of public data on population, financial, and housing
paired with a polygon feature class and two raster datasets (Table 1). The Hurricane Sandy surge
boundary raster used observed flood inundation levels to estimate average depths per census
tract. Impact zones were classified based on depth ranges (Flanagan et al. 2011) and grouped
with average ground elevation heights from the digital elevation model (DEM) for a physical
evaluation (Figure 5). These datasets were merged with the tabular data during analysis to
20

determine the impact zones' population, financial, and housing characteristics before and after
Hurricane Sandy.
Table 1: Data dictionary of input spatial and tabular datasets for the study analysis.
3.1.1. Hurricane Sandy Surge Boundary
As part of FEMA's Natural Hazards Risk Assessment Program (NHRAP), the FEMA
Modeling Task Force (MOTF) generated a storm surge extent with flood depth grid from
Hurricane Sandy data. They created clipped 3-meter DEMs for the state affected: Connecticut,
New Jersey, New York state, and Rhode Island. The United States Geological Survey (USGS)
recorded that Hurricane Sandy hit the Northeast Coast region on October 22, 2012, and did not
dissipate until November 2, 2012. Surge inundation extent was created using the USGS's High-
Water Marks and Storm Surge Sensor data through February 14, 2013. Next, the interpolated
water surface elevation was subtracted from the most recent DEM. For the case of New Jersey,
there was a LiDAR data gap for the state's southwest region, which led to missing data in the
final DEM. Figure 4 shows the boundary extent for New Jersey with the missing data in Salem
Dataset Type Description Source
ACS 5-Year
Estimates
Table
New Jersey's population, financial, and housing
characteristics from 2006 – 2010 and 2014 –
2018 were compared to economic changes
before and after Hurricane Sandy.
US Census
Bureau
Census Tracts
Polygon
Feature
Class
Throughout the study, TIGER/Line census tract
2010 boundaries were used primarily for joining
2010 and 2018 ACS tables.
US Census
Bureau
Hurricane Sandy 3-
meter Surge
Boundary in New
Jersey
Raster
Storm surge extent and depth grid for Hurricane
Sandy layer was used to calculate average flood
inundation per census tract to classify impact
zones.
US FEMA
Modeling Task
Force (MOTF)
New Jersey 10-foot
DEM
Raster
Statewide 10-foot resolution DEM developed
from Light Detection and Ranging (LiDAR)
surveys. It was used to calculate average
elevation and slope for different zones of impact
collection for New Jersey.
New Jersey
Department of
Environmental
Protection
(NJDEP)
21

County, which was considered when estimating census tracts of that region. The figure also
indicates inundation levels from about 0 to 19 feet.

Figure 4. New Jersey's Hurricane Sandy Surge Boundary.
3.1.2. ACS 5-Year Estimates and Accuracy
The US Census Bureau conducts an annual survey known as ACS, in which household
data, such as social, economic, housing, and demographic information, is recorded from a sample
and estimated for a community (Fuller 2018). The survey poses more detailed questions than the
Decennial Census of Population and Housing to give communities current information to plan
investments and services (Parmenter and Lau 2013). For this study, 5-year estimates were
22

analyzed due to the data's availability in low to high population sizes, which can be seen in
census tracts for New Jersey. Consecutively, this can increase the accuracy of the data in
question. Due to the nature of the survey's sampling methodology, each estimate has an
associated Margin of Error (MOE) that equates to the possible variation of the estimated value
(Fuller 2018).
For this study, each of the 16 variables from the ACS Estimates was gathered for New
Jersey at the census tract level and applied to the study's population, income, and housing
evaluations (Appendix A). According to the impact group, the evaluations compared ACS
surveys 2010 and 2018. Due to the aggregation of estimates, the MOE data was also aggregated
using an equation that is the square root of the sum of each MOE estimate multiplied by itself
(Equation 1).
Equation 1. Aggregated Margin of Error Equation.
𝑀𝑂 𝐸 𝑆 𝑢𝑚 =
√
𝑀𝑂 𝐸 𝑒 𝑠 𝑡 1
2
+ 𝑀𝑂 𝐸 𝑒 𝑠 𝑡 2
2
…
The MOE uses the Census Bureau Standard of 90% confidence level, where the lower the
number, the more reliable the data is. To further test the level of reliability of an estimate, each
impact zone’s variable had an associated CV calculated. Aggregated CV used an equation of
dividing the aggregated MOE by 1.645 and dividing that result by the estimate, expressed as a
percentage (Equation 2) (Parmenter and Lau 2013).
Equation 2. Coefficient of Variation for Aggregated Margin of Error Equation.
𝐶𝑉 =
( 𝑀𝑂 𝐸 𝑆 𝑢𝑚 /1 . 645 )
𝐸 𝑠𝑡 𝑖 𝑚𝑎 𝑡 𝑒 ∗ 100
Like MOE, the higher the number, the less reliable the data. When reviewing CV for this
study, reliability will be considered following Parmenter and Lau (2013): less than 15% is high
23

reliability, between 15-30% is moderate reliability, and over 30% is low reliability. ACS Census
Tracts with no population were excluded from the study to maintain the most accurate data.
Also, the final estimations did not include estimates or MOEs with asterisks or dashes due to
these symbolizing too few observations in the source data.
3.2. Research Design
Disaster-induced economic change was assessed using three primary evaluations:
population, income, and housing. The process employed spatial and tabular data that compared
ACS variables before and after Hurricane Sandy. Also, patterns in the value differences were
explored to test clustering and relationships throughout the study area using Hot Spot Analysis
(Getis-Ord Gi*) and Spatial Autocorrelation (Global Moran's I). Before evaluations took place,
the study area was first consolidated to only review New Jersey census tracts with a population
from the 2010 ACS 5-Year Estimate Total Population data table. The tracts with no population
were eliminated from the analysis tables and feature class polygon. The scrubbed census tract
polygon was applied to the impact zone classification using Hurricane Sandy depth grids to
calculate average flood inundation per census tract.
3.2.1. Impact Zone Classification
New Jersey census tracts were allocated according to Hurricane Sandy's surge boundary
using raster and vector data to spatial estimate mean flood inundation depths per census tract.
The analysis utilized ArcGIS Pro's Zonal Statistics as Table tool to calculate the mean depth of
Hurricane Sandy surge raster within each census tract given. The output table from this analysis
was joined to the original 2010 feature class polygon for all future evaluations.
Impact zones were defined by four inundation depths and approximate levels of damage:
None (NIZ) (0 ft), Minor (MIZ) (0-2 ft), Serious (SrIZ) (2-4 ft), and Severe (SvIZ) (>4 ft)
24

(McCarthy et al. 2006). Census tracts with a mean inundation depth evaluated each impact zone
within these ranges. For instance, when average mean land elevation was calculated for the
severely impacted zones, only census tracts with a mean inundation depth above 4-feet were
included in the analysis.
3.2.2. Physical Evaluation
Like the impact zone classification, the census tracts were first allocated according to
mean land elevation and slope for the physical evaluation. First, ArcGIS Pro's Zonal Statistics as
Table tool calculated average elevation grids per census tract using the New Jersey 10-foot DEM
(Figure 5). ArcGIS Pro's Slope tool utilized New Jersey 10-foot DEM to estimate percent raise
with the output of a new raster to calculate slope inclination. Mean, minimum, and maximum
percent slope per census tract was computed using the Zonal Statistics as Table tool. Elevation
and slope estimations were assigned to each census tract, and the output tables were joined to the
2010 census tract polygon containing impact zone classifications. The completed feature class
and joined tables were then exported using ArcGIS Pro's Table to Excel.
The remainder of the physical evaluation was conducted in Microsoft Excel by sorting
and grouping census tract estimations by impact zone. The census tracts in each impact zone
were calculated using the COUNT function in a separate calculations tab. Average mean
elevation, average mean slope, average minimum slope, and average maximum slope were also
calculated for each impact zone. Each category used the AVERAGE function in the calculations
tab to calculate the average result for the census tracts within the given impact zone.
25

Figure 5. Surface Elevation of New Jersey.
3.2.3. Comparison of Years Before and After Disaster
2010 and 2018 data profiles from ACS 5-Year Estimates were used to compare recovery
disparities among Hurricane Sandy impact zones. According to the year range, each evaluation
used associated data profiles (Appendix A) applied to two copies of classified census tracts in
ArcGIS Pro. The 2006 – 2010 ACS data were analyzed first, followed by the 2014 – 2018 ACS
data. Each updated feature class was exported as an Excel spreadsheet and added as a new tab in
the Excel file created in the physical evaluation, where sorting and grouping were used to
calculate the impact zones' results. Calculations from before and after Hurricane Sandy were
26

compared using percent change (Equation 3) by subtracting the 2018 result from the 2010 result
and dividing that outcome by the 2010 result.
Equation 3. Percent Change Equation.
𝑐 =
𝑥 2
− 𝑥 1
𝑥 1
∗ 100
CV was calculated for every ACS variable to ensure transparent reliability by aggregating
each census tracts’ MOE by impact zone to estimate the final CV (Fuller 2018). Also, spatial
statistics, such as clustering and regression, were performed to test the relationship of the
outcomes on a census tract level.
3.2.3.1. Population Evaluation
In ArcGIS Pro, the census tract size was calculated by adding a field named
"Area_sqmil" and using Calculate Geometry to estimate the area in square miles of each
polygon. The 2010 and 2018 ACS "Total Population" data profiles were joined to two separate
copies of the updated and classified census tract feature class using Census Tract ID. Each
feature class copy was exported using Table to Excel and added to the master Excel file as new
tabs.
Total population, total population percentage, and population density were added to the
calculation tab of the spreadsheet for each year range. The total population was calculated using
the SUM function of the people in all census tracts in each impact zone. Total population
percentage was calculated using the total population result and dividing that by the sum of all the
census tract population and multiplying by 100 for percentage. Population Density was
calculated using the total population result and dividing that by the sum of the zone's square mile
area. The Total Population variable from each year's data profile had a calculated CV using the
27

final CV equation. Each zonal result from 2010 was compared to 2018 to calculate the percent
change.
3.2.3.2. Income Evaluation
In ArcGIS Pro, the 2010 and 2018 ACS "Median Income in The Past 12 Months" and
"Poverty Status in The Past 12 Months by Sex by Age" data profiles were joined to two separate
copies of the classified census tract feature class using Census Tract ID. Each feature class copy
was exported using Table to Excel and added to the master Excel file as new tabs. When
comparing dollar value data from 2010 to 2018, the data were not normalized across data
profiles. Therefore, the income evaluation compared data horizontally at its inflation-adjusted
dollar amount for the year presented.
Average median household income, lowest median household income, highest median
household income, and persons below poverty percentage were added to the calculation tab of
the spreadsheet for each year range. Average median household income was calculated using the
AVERAGE function on the Median Household Income variable across each impact zone. The
lowest median household income was calculated using the MIN function on the Median
Household Income variable across each impact zone. The highest median household income was
calculated using the MAX function on the Median Household Income variable across each
impact zone. The Median Household Income and Persons Below Poverty Estimate variables
from each year's data profile had a calculated CV using the final CV equation. Each zonal result
from 2010 was compared to 2018 to calculate the percent change.
Household income classes were estimated using the Pew Research Center standards and
the 2010 and 2018 ACS "Income in The Past 12 Months" data profiles. According to a 2016
report, the national middle-class household income range was between $45,200 and $135,600
28

(Kochhar 2018). With best efforts to meet this range, this study categorized the middle-class
between $50,000 and $149,999. Values below that range were considered lower-class, and those
above that range were considered upper-class. The 2010 and 2018 ACS "Income in The Past 12
Months" data profiles were first edited in Excel to create a total percentage of households per
income class per census tract. Each variable within the range of the income class was added to a
column labeled low, middle, or upper class. Using Census Tract ID, this edited table joined two
separate copies of the classified census tract feature class in ArcGIS Pro. Each feature class copy
was exported using Table to Excel and added to the master Excel file as new tabs.
Lower-income households, middle-income households, and upper-income households
were added to the calculation tab of the spreadsheet for each year range. Lower-income
households were calculated by adding the total number of households with less than $50,000
annual income by impact zone and dividing it by the total number of households by impact zone.
Middle-income households were calculated by adding the total number of households between
$50,000 and $150,000 annual income by impact zone and dividing it by the total number of
households by impact zone. Upper-income households were calculated by adding the total
number of households with more than $150,000 annual income by impact zone and dividing it by
the total number of households by impact zone. The Estimated Household Income variables from
each year's data profile had a calculated CV using the final CV equation. Each zonal result from
2010 was compared to 2018 to calculate the percent change.
3.2.3.3. Housing Evaluation
In ArcGIS Pro, the 2010 and 2018 ACS "Households and Families" and "Vacancy
Status" data profiles were joined to two separate copies of the classified census tract feature class
29

using Census Tract ID. Each feature class copy was exported using Table to Excel and added to
the master Excel file as new tabs.
Owner-occupied housing units, renter-occupied housing units, and vacant housing units
were added to the calculation tab of the spreadsheet for each year range. Owner-occupied
housing units were calculated as a percentage by adding the total number of owned housing units
in the impact zones and dividing it by the total number of all housing units by impact zone.
Renter-occupied housing units were calculated as a percentage by adding the total number of
rented housing units in the impact zones and dividing it by the total number of all housing units
by impact zone. Vacant housing units were calculated as a percentage by adding the total number
of vacant housing units in the impact zones and dividing it by the total number of housing units
by impact zone. The Owner-Occupied, Renter-Occupied, and Vacant Housing Unit variables
from each year's data profile had a calculated CV using the final CV equation. Each zonal result
from 2010 was compared to 2018 to calculate the percent change.
3.2.3.4. Spatial Statistics
Before statistics were applied, the value difference of each variable was calculated to
estimate socio-economic change by census tract from 2010 to 2018. The value differences were
calculated in a new CSV file by subtracting 2010 data from 2018 data for population, population
density, median household income, Persons below poverty, lower-income households, middle-
income households, upper-income households, owner-occupied housing, renter-occupied
housing, and vacant housing units. In ArcGIS Pro, the table was joined to a copy of the updated
and classified 2010 census tract feature class using Census Tract ID. To estimate the occurrence
of clustering in the study area, Incremental Spatial Autocorrelation was first implemented to
determine the peak distance of statistically significant clustering. Using the optimal distance in
30

the fixed distance band option, the Hot Spot Analysis (Getis-Ord Gi*) tool was deployed for each
variable difference. Clustering of high and low values was further analyzed with a final spatial
statistic to test the distribution and reliability of the data. The Spatial Autocorrelation (Global
Moran's I) tool was used to indicate whether the features were spatially correlated, while
randomly distributed components favored the null hypotheses. A positive Moran's I value rejects
the null hypothesis by expressing a tendency toward clustering. Spatial significance was rated
according to the z-score value, where the higher the number was, the more statistically
significant the spatial relationships were.

31

Chapter 4 Results
The study aimed to examine the socio-economic characteristics of existing coastal communities
by developing an evaluation to understand the relationship between recovery and physical
exposure from a severe flood event. The impact zone classification indicated that, for the 1,999
census tracts analyzed, more than half of the population resided in NIZ. The physical evaluation
revealed that New Jersey's elevation and slope were relatively uniform except for the dramatic
elevation increase in the northwestern NIZ, accounting for four times the average mean height of
the other zones. Overall, the community's transformation from before to after Hurricane Sandy
was not overtly apparent; across all the zones, the population increased, the average median
household income increased, persons below poverty increased, and owner-occupied housing
decreased. Some distinct findings were that the tract with the lowest median household income
in the SrIZ saw an increase of 66% in median household income, and the tracts with the highest
median household income in the NIZ and SvIZ increased almost 100%. The MIZ had the softest
increase in median household income and the highest increase in renter-occupied housing
compared to the other impact zones. The CV for all the aggregated ACS data proved reliable,
with the highest unreliability being 9%. However, the non-aggregated ACS data for the lowest
and highest median household income proved to be exceptionally high. The complete results of
New Jersey's population, financial, and housing data change are outlined in this chapter.
4.1.1. Impact Zone Classification
Data scrubbing eliminated five of the 2010 census tract polygons with nominal
population size and margin of error, leaving a total of 1,999 census tracts used in the study. The
ArcGIS Spatial Analyst Tool Zonal Statistics as Table calculated mean flood inundation of the
Hurricane Sandy Surge Boundary depth grids across the census tract polygons. Figure 6 shows

32

the impact zones were classified according to impact zones of None (NIZ) (0 ft), Minor (MIZ)
(0-2 ft), Serious (SrIZ) (2-4 ft), and Severe (SvIZ) (>4 ft) (McCarthy et al. 2006). The totals for
the study area include the NIZ with 1,329 tracts, followed by SvIZ with 363, SrIZ with 194, and
MIZ with 83 census tracts.

Figure 6. Hurricane Sandy Surge Impact Zones of New Jersey.
4.1.2. Physical Evaluation
Before comparing 2010 and 2018 ACS data, the study area's land elevation and slope
were examined using the New Jersey 10-foot DEM concerning 2010 census tracts. The analysis
showed that around two-thirds of the census tracts examined remained in the NIZ, with the

33

smallest sample size residing in the MIZ (Table 2). The NIZ exhibited the highest average mean
elevation of 194 feet, average mean slope of 2.9 percent rise, and average maximum slope of
11.1 percent rise. Impact zones Minor, Serious, and Severe had similar results for average mean
elevation, average mean slope, and average minimum slope. SvIZ did show the second highest
average maximum slope; however, it was closer to the MIZ and SrIZ than NIZ.
Table 2. Physical Evaluation of Elevation and Slope per Impact Zone.
Impact Zone Census
Tracts
Average
Mean
Elevation
(Feet)
Average Mean
Slope (Percent
Rise)
Average
Minimum
Slope
(Percent Rise)
Average
Maximum
Slope (Percent
Rise)
None (0 ft)** 1329* 194* 2.9* 0.2* 11.1*
Minor (0-2 ft) 83 41 1.8 0.2* 5.5
Serious (2-4 ft) 194 33 1.5 0.1 5.2
Severe (>4 ft) 393 30 1.6 0.0 7.2
* Result(s) with the highest value in the dataset.
** Impact Zone(s) in the dataset with the largest difference in results from the evaluation.

The zonal statistics spatial analyst estimated the highest mean elevation and slope in the
state's northwestern region, coinciding with most NIZ (Figure 7 and 8). The distribution of mean
elevation throughout the census tracts was displayed in Figure 7, highlighting the difference in
incline from 0.98 to 1,200 feet. While Figure 8 showed a slightly more distributed mean slope
with the highest percentage rise in the northwest and some in the central-eastern region of the
state. Overall, the physical evaluation showed that New Jersey's elevation and the slope were
relatively uniform throughout most of the study area, with the highest in the northwestern region.

34

Figure 7. Average Mean Elevation per Census Tract in New
Jersey.

Figure 8. Average Mean Slope per Census Tract in New Jersey.

35

4.1.3. Comparison of Years Before and After Disaster
In comparing 2010 and 2018 data profiles from ACS 5-Year Estimates, the tabular data
was linked to impact zone classified 2010 census tract data to evaluate population, income, and
housing. In addition, the reliability of the data was tested according to MOE aggregation and the
CV of each variable in the impact zones. The ACS data used in the study proved to be highly
reliable, with an average of 1.67% and the least reliable variable being 9% from 2018 Median
Household Income in the MIZ (Table 4). According to the CV results, the most reliable ACS
data examined was total population, and the least reliable was median household income.
4.1.3.1. Population Evaluation
The population evaluation analyzed the total population, population percentage, and
population density of census tracts by impact zone using 2010 and 2018 ACS data profiles. The
highest total population resided in the NIZ, with more than 6 million residents with the lowest
total population in the MIZ (Table 3). In comparing data years 2010 and 2018, population
distribution remained about the same, with more than half of the total population in the NIZ
followed by 18% in SvIZ.
Table 3. Population Evaluation per Impact Zone for ACS Years 2006 – 2010 and 2014 – 2018.
ACS 5-year
Dataset
Impact Zone Population Total
Population
Percentage
Population Density
(Population per
square mile)
CV for
Population
Estimate
2006-2010 None (0 ft)** 6,031,482* 69%* 1,250 0%
Minor (0-2 ft) 333,117 4% 1,729* 1%
Serious (2-4 ft) 769,335 9% 1,551 0%
Severe (>4 ft)** 1,587,489 18% 687 0%
2014-2018 None (0 ft)** 6,119,636* 69%* 1,268 0%
Minor (0-2 ft) 345,973 4% 1,797* 1%
Serious (2-4 ft) 779,045 9% 1,566 0%
Severe (>4 ft)** 1,636,028 18% 708 0%
* Result(s) with the highest value in the dataset.
** Impact Zone(s) in the dataset with the largest difference in results from the evaluation.

36

The population steadily increased across all impact zones, with the most percent increase
in the MIZ of 4% due to its already low population, although growth rates were roughly the same
across all zones (Figure 9). The SvIZ also showed a population of over 1.5 million with a percent
change of 3%. The total population across the census tracts from 2010 and 2018 ACS data
profile estimates only increased by 159,259 individuals.

Figure 9. Population Change Graph, 2006 – 2010 and 2014 – 2018.
At a glance, most of the highest density areas were in the northeastern and south to the
central-western region of the state, which intersects the Hurricane Sandy Surge boundary (Figure
10). In 2010, the remainder of the state, although still dense, had a population density below
5,000 people per square mile. Table 3 examination by impact zone displayed a similar population
density in 2010 between 1,250 and 1,729 people per square mile in impact zones None, Minor,
and Serious. Although the densest census tracts in the state were in the SvIZs, the overall
population density for this zone was the lowest, with only 687 people per square mile. Most of
this zone had lower population density with more extensive census tracts; therefore, the low
+1%
+4%
+1%
+3%
0
1,000,000
2,000,000
3,000,000
4,000,000
5,000,000
6,000,000
7,000,000
2006-2010
2014-2018
2006-2010
2014-2018
2006-2010
2014-2018
2006-2010
2014-2018
None (0ft) Minor (0-2ft) Serious (2-4ft) Severe (>4ft)

37

population regions overpowered the pockets of high density. Population density did not increase
substantially in 2018 due to the steady overall increase in population across all impact zones.

Figure 10. 2010 Population Density per Census Tract in Hurricane Sandy Surge Boundary.
4.1.3.2. Income Evaluation
The income evaluation analyzed the median household income, persons below poverty,
and income classes of census tracts by impact zone using 2010 and 2018 ACS data profiles. In
2010, the average median household income throughout the impact zones ranged between
$62,817 and $77,906 (Table 4). However, in 2018 the average median household income
broadened its range to $67,838 and $89,752. Comparing the two data profile years across impact

38

zones, the results showed that average median household income experienced a 15% change
increase in NIZ, MIZ, and SrIZ, with a 16% change in the SvIZ (Figure 11).
In 2010, the lowest median household income for all impact zones ranged from $9,393 to
$13,624, with the lowest in the SrIZ and the highest in the NIZ. Conversely, in 2018, the lowest
median household income range became more rigid, ranging between $12,443 to $15,579 (Table
4). The lowest impact zone switched to the MIZ, while the SrIZ changed from the lowest in 2010
to the highest in 2018. Comparing the two data profile years, the results showed that the lowest
median household income increased overall but at different levels of percent change (Figure 12).
The SrIZ grew at 66%, followed by MIZ at 29% and SvIZ at 25%. The NIZ experienced a minor
change with an increase of 8%, showing that this zone's lowest median household income
remained comparatively constant from 2010 to 2018. When reviewing the reliability of the
individual census tract where median income originated, all the 2018 ACS data had a standard
deviation greater than the mean. The highest CV in the 2010 ACS data was 21%, making this
census tract less reliable than the rest of the impact zones but still purposeful. The other CV
values for the 2018 ACS data were too high to be considered reliable.
The highest median household income in 2010 ranged from $131,298 in the SrIZ to
$238,162 in the NIZ. While the highest median household income in 2018 ranged from $171,635
to $441,283 in the same impact zones as 2010 (Table 4). The highest median household income
distribution remained the same in 2010 and 2018; however, impact zones None and Severe
considerably increased. The NIZ displayed an increase of 85%, and the SvIZ revealed a 90%
increase, while MIZ increased 15% and SrIZ experienced a 31% increase (Figure 13). The
impact zones with the original highest median household income, None and Severe, increased
significantly while the other zones, Minor and Serious, increased marginally by comparison.

39

When reviewing the reliability of the individual census tract where median income originated,
most of the ACS data was unfavorable. The CV values for the 2010 ACS data ranged from 7% to
71%, and 2018 ACS data ranged from 61% to 71%. The U.S. Census case studies consider above
30% CV to be low reliability; therefore, all the 2018 ACS data in this evaluation cannot be
considered trustworthy. However, most of the impact zones from the 2010 ACS data were
reliable except for MIZ.
In 2010, the population percentage considered below the poverty level in each impact
zone ranged from 8% to 12%. The persons below poverty percentage from 2018 resulted in a
range more uniform, with impact zones None and Severe at 10% and Minor and Serious at 13%
(Table 4). Poverty increased overall, with the most remarkable percent change of 17% in the
MIZ (Figure 14). Considering the 2019 ACS 5-Year Estimates revealed the national poverty
level to be 13.4%, New Jersey's general poverty level remains low (US Census Bureau).

40

Table 4. Median Household Income Evaluation per Impact Zone for ACS Years 2006 – 2010 and 2014 – 2018.
ACS 5-
year
Dataset
Impact Zone Average
Median
Household
Income
CV for
Median
Household
Income
Lowest
Median
Household
Income
CV for
Lowest
Median
Household
Income
Highest
Median
Household
Income
CV for
Highest
Median
Household
Income
Persons
Below
Poverty
(Percentage
of
Population)
CV for
Persons
Below
Poverty
2006- None (0 ft)** $77,906* 0% $13,624* 13% $238,162* 11% 8% 1%
2010 Minor (0-2 ft) $62,817 3% $9,631 13% $162,500 71%* 11% 4%
Serious (2-4ft)** $59,160 1% $9,393 15% $131,298 7% 12%* 3%
Severe (>4 ft) $70,212 1% $12,210 21%* $214,323 8% 9% 2%
2014- None (0 ft) $89,752* 2% $14,729 125%* $441,283* 61% 10% 1%
2018 Minor (0-2 ft) $72,430 9% $12,443 111% $187,121 71%* 13%* 3%
Serious (2-4ft)** $67,838 6% $15,579* 107% $171,635 68% 13%* 2%
Severe (>4 ft) $81,163 4% $15,243 112% $407,346 61% 10% 2%
* Result(s) with the highest value in the dataset.
** Impact Zone(s) in the dataset with the largest difference in results from the evaluation.

41

Figure 11. Average Median Household Income Graph, 2006-
2010 and 2014-2018.

Figure 12. Lowest Median Household Income Graph, 2006-
2010 and 2014-2018.

Figure 13. Highest Median Household Income Graph, 2006-
2010 and 2014-2018.

Figure 14. Persons Below Poverty Percentage Graph, 2006-
2010 and 2014-2018.
+15%
+15%
+15%
+16%
$0
$10,000
$20,000
$30,000
$40,000
$50,000
$60,000
$70,000
$80,000
$90,000
$100,000
2006-2010
2014-2018
2006-2010
2014-2018
2006-2010
2014-2018
2006-2010
2014-2018
None (0ft) Minor (0-2ft) Serious (2-4ft) Severe (>4ft)
+8%
+29%
+66%
+25%
$0
$2,000
$4,000
$6,000
$8,000
$10,000
$12,000
$14,000
$16,000
$18,000
2006-2010
2014-2018
2006-2010
2014-2018
2006-2010
2014-2018
2006-2010
2014-2018
None (0ft) Minor (0-2ft) Serious (2-4ft) Severe (>4ft)
+85%
+15%
+31%
+90%
$0
$50,000
$100,000
$150,000
$200,000
$250,000
$300,000
$350,000
$400,000
$450,000
$500,000
2006-2010
2014-2018
2006-2010
2014-2018
2006-2010
2014-2018
2006-2010
2014-2018
None (0ft) Minor (0-2ft) Serious (2-4ft) Severe (>4ft)
+14%
+17% +12%
+14%
0
10
20
30
40
50
60
70
80
90
100
2006-2010
2014-2018
2006-2010
2014-2018
2006-2010
2014-2018
2006-2010
2014-2018
None (0ft) Minor (0-2ft) Serious (2-4ft) Severe (>4ft)

42

Household income classes in this study are defined as lower (less than $50,000 annual
income), middle ($50,000 - $150,000 annual income), and upper (more than $150,000 annual
income). The distribution of household income classes in 2010 and 2018 proved to be similar
throughout the impact zones, with most of the households in the middle-income class (Table 5).
In 2010, the highest percentage of upper-income households and the lowest percentage of lower-
income households resided in the NIZ. Also, the SrIZ was the lowest percentage of upper-
income households and the highest percentage of lower-income households in 2010. This
distribution remained true for the 2018 results.
When comparing 2010 and 2018, there was a percent change decrease of 9% throughout
the impact zones for lower-income households; however, the SvIZ showed a considerable
reduction of 12% (Figure 15). Middle-income households also showed a consistent decrease
among impact zones, with a decline of 7% in the NIZ (Figure 16). Upper-income households
showed the most variation in percent change with increased values up to 64% in the SrIZ and an
increased 34% in the NIZ (Figure 17). The inundated zones have a higher percent change
compared to NIZ, but there is not a consistent increase towards deeper flood depth.

43

Table 5. Income Classes Evaluation per Impact Zone for ACS Years 2006 – 2010 and 2014 – 2018.

ACS 5-year
Dataset
Impact Zone Lower-
Income
Households
Percentage
(Less than
$50,000
Annual
Income)
CV for
Lower-
Income
Households
Middle-Income
Households
Percentage
($50,000 -
$150,000
Annual
Income)
CV for
Middle-
Income
Households
Upper-Income
Households
Percentage
(More than
$150,000
Annual
Income)
CV for
Upper-
Income
Households
2006-2010 None (0 ft) 34% 1% 48%* 0% 18%* 1%
Minor (0-2 ft) 43% 4% 46% 3% 11% 8%
Serious (2-4 ft)** 44%* 1% 47% 1% 9% 3%
Severe (>4 ft) 37% 1% 48%* 1% 14% 2%
2014-2018 None (0 ft)** 31% 1% 44% 0% 24%* 1%
Minor (0-2 ft) 39% 4% 44% 3% 17% 4%
Serious (2-4 ft) 40%* 1% 45% 1% 15% 2%
Severe (>4 ft)** 33% 1% 46%* 1% 21% 1%
* Result(s) with the highest value in the dataset.
** Impact Zone(s) in the dataset with the largest difference in results from the evaluation.

44

Figure 15. Lower-Income Households Percentage Graph, 2006 – 2010 and 2014 – 2018.

Figure 16. Middle-Income Households Percentage Graph, 2006 – 2010 and 2014 – 2018.

Figure 17. Upper-Income Households Percentage Graph, 2006 – 2010 and 2014 – 2018.
-9%
-9% -9%
-12%
0
10
20
30
40
50
60
70
80
90
100
2006-2010
2014-2018
2006-2010
2014-2018
2006-2010
2014-2018
2006-2010
2014-2018
None (0ft) Minor (0-2ft) Serious (2-4ft) Severe (>4ft)
-7% -5% -4%
-5%
0
10
20
30
40
50
60
70
80
90
100
2006-2010
2014-2018
2006-2010
2014-2018
2006-2010
2014-2018
2006-2010
2014-2018
None (0ft) Minor (0-2ft) Serious (2-4ft) Severe (>4ft)
+37%
+55%
+64%
+47%
0
10
20
30
40
50
60
70
80
90
100
2006-2010
2014-2018
2006-2010
2014-2018
2006-2010
2014-2018
2006-2010
2014-2018
None (0ft) Minor (0-2ft) Serious (2-4ft) Severe (>4ft)

45

4.1.3.3. Housing Evaluation
The housing evaluation analyzed the owner-occupied, renter-occupied, and vacant
housing units of census tracts by impact zone using 2010 and 2018 ACS data profiles. In 2010,
owner-occupied housing units were similar among all the impact zones ranging from 58% to
69%. The corresponding renter-occupied housing units were similar across impact zones and
ranged from 31% to 42% (Table 6). The peak percentage of owner-occupied units was in the
NIZ, and the lowest rate was the MIZ. In 2018, housing units exhibited slightly more variation
among impact zones with 53% and 66% of owner-occupied ranges, respectively, and renter-
occupied ranges between 34% and 47%. The highest and lowest percentage for the impact zone
remained the same for 2018.
Throughout the comparison of 2010 and 2018 data, overall owner-occupied housing units
decreased, and renter-occupied housing units increased (Figure 18 and 19). Impact zones None,
Serious, and Severe decreased by 4% and 5%, while Minor decreased 9% (Figure 18). The
corresponding renter-occupied housing units increased by 9% in impact zones None and Severe.
The SrIZ had the lowest increase at 6%, and MIZs had the highest at 12% (Figure 19). The data
shows that the most occupied housing change occurred in the MIZ.
In 2010, vacant housing units were predominately in the SvIZ, with 22% of housing units
vacant in this zone. A similar distribution of vacant housing units was seen in 2018, with 24%
vacancies in the SvIZ (Table 6). Housing vacancy trended towards the most impacted zones.
Vacancies increased in all impact zones from 2010 to 2018, with the highest percent change of
12% in the SrIZ and the lowest percentage change of 7% in the SvIZ (Figure 20).

46

Table 6. Housing Evaluation per Impact Zone for ACS Years 2006 – 2010 and 2014 – 2018.
ACS 5-year
Dataset
Impact Zone Owner-
Occupied
Housing
Units
CV for
Owner-
Occupied
Housing
Units
Renter-
Occupied
Housing
Units
CV for
Renter-
Occupied
Housing
Units
Vacant
Housing
Units
CV for
Vacant
Housing
Units
2006-2010 None (0 ft) 69%* 0% 31% 0% 7% 1%
Minor (0-2 ft) 58% 2% 42%* 2% 8% 4%
Serious (2-4 ft) 60% 1% 40% 1% 17% 2%
Severe (>4 ft)** 66% 0% 34% 1% 22%* 1%
2014-2018 None (0 ft) 66%* 0% 34% 0% 8% 1%
Minor (0-2 ft) 53% 1% 47%* 2% 9% 4%
Serious (2-4 ft) 57% 1% 43% 1% 19% 1%
Severe (>4 ft)** 63% 0% 37% 1% 24%* 1%
* Result(s) with the highest value in the dataset.
** Impact Zone(s) in the dataset with the largest difference in results from the evaluation.

47

Figure 18. Percentage of Owner-Occupied Housing Units Graph, 2006-2010 and 2014-2018.

Figure 19. Percentage of Renter-Occupied Housing Units Graph, 2006-2010 and 2014-2018.

Figure 20. Percentage of Vacant Housing Units Graph, 2006-2010 and 2014-2018.
-4%
-9%
-4%
-5%
0
10
20
30
40
50
60
70
80
90
100
2006-2010
2014-2018
2006-2010
2014-2018
2006-2010
2014-2018
2006-2010
2014-2018
None (0ft) Minor (0-2ft) Serious (2-4ft) Severe (>4ft)
+9%
+12%
+6%
+9%
0
10
20
30
40
50
60
70
80
90
100
2006-2010
2014-2018
2006-2010
2014-2018
2006-2010
2014-2018
2006-2010
2014-2018
None (0ft) Minor (0-2ft) Serious (2-4ft) Severe (>4ft)
11%
10%
12%
7%
0
10
20
30
40
50
60
70
80
90
100
2006-2010
2014-2018
2006-2010
2014-2018
2006-2010
2014-2018
2006-2010
2014-2018
None (0ft) Minor (0-2ft) Serious (2-4ft) Severe (>4ft)

48

4.1.3.4. Spatial Statistics
The statistical analysis evaluated the value changes from 2010 to 2018 for population,
population density, median household income, people below poverty, lower-income households,
middle-income households, upper-income households, owner-occupied housing, renter-occupied
housing, and vacant housing units. The Incremental Spatial Autocorrelation tool calculated
statistically significant z-score peaks at 20,542 meters and 25,701 meters to indicate distances
where clustering of census tracts is most pronounced. Using an estimated distance of statistically
significant clustering, 21,000-meter fixed distance band was used in the Hot Spot Analyses.
The Hot Spot Analysis (Getis-Ord Gi*) tool estimated the clustering census tracts of
highest and lowest values for the variables' value change. Clustering of population change from
2010 to 2018 was documented in Figure 21. The northeastern region to the state's central region
experienced high to moderate clustering of high values. While clustering of low values was less
concentrated, expressing high to low clustering in the northwest, south, and central east regions.
Population Density was a much more concentrated value change, with high clustering of high
values in the northeastern region and high clustering of low values in the central west area
(Figure 22). The highest number of census tracts for both population cluster analyses were
insignificant.

49

Figure 21. Hot Spot Analysis of Population Difference from
2010 – 2018.

Figure 22. Hot Spot Analysis of Population Density Difference
from 2010 – 2018.

50

Clustering of median household income change from 2010 to 2018 was documented in
Figure 23. The crescent-shaped high to moderate clustering of high values were in New Jersey's
northeast region and did not interact much with the Hurricane Sandy storm surge boundary.
However, most of the high to low clustering of low values were along the storm surge boundary
from the northeast to the southwest regions. The gap of insignificant clustering in the
northeastern part of median household income was also expressed as a moderate to low
clustering of high values of persons below poverty in Figure 24. Low-value clustering is not well
displayed in the Hot Spot Analysis due to the overall increase in poverty levels from 2010 to
2018. Most of the persons below poverty clusters of change were along the coastline; however, it
is not as significant as median household income change. The highest number of census tracts for
median household income and persons below poverty cluster analyses were insignificant.
Clustering of lower-income household change from 2010 to 2018 was documented in
Figure 25. High to low clustering of high values were observed in the state's central region, with
a few outliers in the north and southeast. The increased number of lower-income households in
these sections did not interact much with the Hurricane Sandy storm surge boundary. However,
the central region appears to branch off from the boundary with decreased clustering. While high
to moderate clustering of low values were expressed solely on the coastline, interacting with the
storm surge boundary. Concentrations of reducing lower-income households were observed in
the northeastern area, central-eastern region, and southern peninsula of the state. The patterns of
lower-income households did not reflect the same patterns shown in the middle- and the upper-
income household trends. The highest number of census tracts for the lower-income households
cluster analysis were insignificant, with most of the tracts in the state's southern section.

51

Figure 23. Hot Spot Analysis of Median Household Income
Difference from 2010 – 2018.

Figure 24. Hot Spot Analysis of Persons Below Poverty
Difference from 2010 – 2018.

52

Figure 25. Hot Spot Analysis of Lower-Income Household Difference from 2010 - 2018.
The Hot Spot Analysis for middle- and upper-income household change from 2010 to
2018 shows related trends (Figure 26 and 27). High clustering of increased middle-income
households in the northeastern area also showed moderate clustering of decreased upper-income
households. While high to moderate clustering of middle-income households in the rest of the
northern to the central region also showed high to moderate clustering of increased upper-income
households. Outliers for middle-income expressed a high-value cluster in the southeast, and
upper-income represented a low-value cluster in the southern coastline. Both income households
showed a moderate to low clustering of low values in the southwestern region.

53

Figure 26. Hot Spot Analysis of Middle-Income Household
Difference from 2010 - 2018.

Figure 27. Hot Spot Analysis of Upper-Income Household
Difference from 2010 - 2018.

54

The Hot Spot Analysis for owner- and renter-occupied housing unit change from 2010 to
2018 shows related trends (Figure 28 and 29). Instead of the clustering showing adverse results
from owner to renter values, the clustering was comparable. Owner-occupied housing units
expressed high clustering of high values in a small area to the southeastern region with scattered
moderate to low clustering in the northeast, central, and southwest areas. High clustering of high
values for renter-occupied housing units was more concentrated to the northeastern and the
central regions, with a small moderate to low high cluster in the southeastern area.
A small cluster of decreasing owner-occupied housing units was observed in the
southwestern and southeastern areas. However, the rest of the low-value clusters were moderate
to low. High and low values of clusters for owner-occupied housing units were scattered
throughout the study area. High clustering of low-value renter-occupied housing units was
mainly in the southern peninsula. The other low value moderate to low clusters in the southwest
and central east areas. There does not appear to be a relation between high clustering and
proximity to the Hurricane Sandy storm surge boundary. The highest number of census tracts for
the owner- and renter-occupied housing unit cluster analysis was insignificant with most of the
tracts in the state's northern section.

55

Figure 28. Hot Spot Analysis of Owner-Occupied Housing
Difference from 2010 – 2018.

Figure 29. Hot Spot Analysis of Renter-Occupied Housing
Difference from 2010 – 2018.

56

There was a clear difference between cluster values and regions for vacant housing unit
change from 2010 to 2018 (Figure 30). High clustering of high values was displayed throughout
most of the southern region and part of the northwestern area. Most of the high clustering of
increased vacancies in the southern region was along the Hurricane Sandy storm boundary. High
clustering of low values was exclusively in the northeastern area, with some moderate to low
clustering in the central area, which all interacted with the storm surge boundary. Vacant housing
unit Hot Spot Analysis was the only clustering analysis that expressed higher low-value census
tracts than insignificant tracts.

Figure 30. Hot Spot Analysis of Vacant Housing Unit Difference from 2010 – 2018.

57

To validate the statistical distribution and reliability of the value changes, the Spatial
Autocorrelation (Global Moran's I) tool tested census tracts' relation to each other. All variables
investigated had a clustered distribution except for persons below poverty, which had a random
distribution (Table 7). The variable with the most significant clustering was vacant housing units
with a z-score above 30. Variables with low z-scores and clustered distribution were owner-
occupied housing units, median household income, and lower-income households. Due to their
low scores, caution was used when interpolating the Hot Spot Analysis.
Table 7. Spatial Autocorrelation (Global Moran's I) of Variable Difference from 2010-2018.
Variable Difference
from 2010 to 2018
Moran's Index Critical Value
(z-score)
Distribution Type
Population 0.024 8.897 Clustered
Population Density 0.018 6.807 Clustered
Median Household Income 0.015 5.583 Clustered
Persons Below Poverty** -0.002 -0.402 Random
Lower-Income Households 0.015 5.615 Clustered
Middle-Income Households 0.030 10.943 Clustered
Upper-Income Households 0.032 11.622 Clustered
Owner-Occupied Housing 0.008 3.097 Clustered
Renter-Occupied Housing 0.026 9.486 Clustered
Vacant Housing Units** 0.083 30.045* Clustered
* Result(s) with the highest value in the dataset.
** Variable(s) in the dataset with the largest difference in results from the analysis.

58

Chapter 5 Discussion and Conclusion
This study analyzed the relationship between New Jersey communities' economic standing and
level of impact after Hurricane Sandy's coastal flooding event by evaluating population, income,
and housing metrics before and after the disaster. The analysis designated impact zones
according to the mean flood inundation depth estimated in each census tract observed. The study
aimed to understand a community's recovery in various inundation levels and whether proximity
to severe flooding results in a notable change. Comprehending this association can assist in
protecting the neighborhoods most likely to struggle after an extreme flood event. The impact
zone classification used to organize results into approximate levels of damage—None (NIZ) (0
ft), Minor (MIZ) (0-2 ft), Serious (SrIZ) (2-4 ft), and Severe (SvIZ) (>4 ft)—by variable and
year. The physical evaluation of elevation, slope, and census tract count was first introduced to
examine the zones' characteristics that may influence comparative assessments. The 2010 and
2018 ACS data profile estimates were assessed for each impact zone's population, income, and
housing evaluations. The following chapter discusses the significance of this study's assessment
findings and limitations and suggested future research.
5.1. Limitations and Future Research
Although the evaluations successfully obtained insight into socio-economic conditions in
New Jersey's census tracts and classified impact zones, there remains room for improvement in
the data and analyses. Notable limitations in data quality, methodology factors, and research
scale decreases the reliability of the study's results. The two data quality limitations identified
were: reliably of ACS data and Hurricane Sandy storm surge data gaps. Next, the three
methodology factor limitations identified were: impact zone classifications, normalization of
dollar-values, and broader vulnerability focus. Finally, the only research scale limitations

59

identified was examining a large study area rather than subregions. These limitations can
produce a more accurate and desirable outcome through further research.
The core tabular data used in the study was a 2010 and 2018 ACS data profile
comparison of population, income, and housing throughout the four classified impact zones.
Although the variables provided adequate knowledge of the study area's conditions before and
after Hurricane Sandy, further research would benefit from a deeper understanding of New
Jersey's vulnerability itself. Highlighting site-specific weaknesses in socio-economic conditions
help to identify valid notable changes. Some data quality drawbacks would benefit from a
comprehensive analysis of the vulnerability of data collected.
In comparing household income from 2010 to 2018 ACS data, income appears to
increase more than they have due to the values not using a constant dollar value. Although the
data was not normalized, the results were not impacted because the compared zones would still
result in the same trends. Future analysis comparing incomes from 2010 and 2018 ACS would
benefit from constant dollar values.
At the census tract level, some ACS data were considered unreliable due to the estimated
variables and range of error applied to each result. If there was not enough sample data collected
for the surveys, the estimated findings would be less reliable. However, grouping data in
classified zones help increase trustworthiness by aggregating MOE and averaging less reliable
data. In this study, census tracts were grouped by impact zones where MOE was aggregated for
most examined variables. Some evaluations, such as lowest and highest median household
income, did not call for grouping of ordinal categories, resulting in magnified CV values. The
spatial statistics assessment also compared data on an individual census tract level during hot
spots analysis. Due to higher uncertainty of non-aggregated results, the hot spot analysis was

60

considered exploratory. Although ACS provides more specialized and accessible data, it is less
reliable than decennial census data. With the new 2020 decennial census reports, New Jersey’s
communities can evaluate population and household trends on a more reliable ten-year range.
Another primary input data used in the study was the Hurricane Sandy storm surge
boundary. Due to the observed data gap in the layer, census tracts in the western region of Salem
County were a limiting factor in the results. The other coastal census tracts categorized as SrIZ
rather than SvIZ might have originated from the overlap of inundation boundary to the census
tract. The flood depth range within a census tract could have varied greatly, especially if the
census tract is large. Therefore, the mean depth would have decreased during zonal statistics if
the census tract had areas with no flood depth. Limitations in the storm surge boundary could be
improved by updating the DEM on which the layer was based.
For this study, the impact zone classifications followed the methodology introduced by
McCarthy et al. (2006) to examine storm recovery in New Orleans post-Hurricane Katrina. Upon
further evaluation, the depth intervals would have benefited from adding another impact zone
due to the maximum depth reaching 19 feet. Through this more comprehensive analysis, the
researcher may find more considerable recovery disparities in areas of over 10 feet. The impact
zone depths may have been too limiting for estimates in a more extensive and diverse landmass,
like New Jersey, rather than New Orleans.
In the hot spot analysis review, there was apparent clustering in certain regions identified
as New York Metropolitan area, Central Raritan River, Philadelphia Metropolitan area, Delaware
Bay region. One major limitation was the scale of the study area due to the lack of a clear
connection between the impact zone and notable change. A more concentrated analysis of the
trends within these clustered regions may further explain the relationship. Overall, it would be

61

best to generate a site-specific impact zone range and clustered communities to evaluate disaster-
influenced recovery disparities.
There is potential for further exploration of a natural disaster's influence on a
community's economic recovery, especially in severe flooding. The likelihood of flood-related
damage will continue to increase as climate change, and SLR continues to grow as a threat to
coastal communities. Hurricane Sandy pressed the importance of protecting vulnerable areas
resulting in government funds given to citizens to ensure fair and equal recovery. Due to rising
sea levels and the risk they pose on an explaining population, FEMA flood hazard areas will
need to be revised. Remapping the zones will assist in locating vulnerable areas and impact
subsidy distribution such as flood insurance and floodplain buyback programs such as the
NJDEP Blue Acres program. However, locating the vulnerable areas before a severe flood event
supports local officials in establishing better safeguards to protect these communities
preemptively.
A community can experience harm in forms other than physical destruction and
economic hardship caused by a disaster. Although the concept was not explored in this study, it
would be noteworthy to explore further New Jersey's recovery of Hurricane Sandy from the lens
of disaster-influenced gentrification. Exploring future research in this subject would include
examining housing prices and cost of living along with socio-economic evaluations. An area's
shift from poor neighborhoods to wealthy, boutique lofts can often be accelerated due to a
disaster. Tracking that likelihood may encourage more affordable housing developments in the
areas marked at risk.

62

5.2. Conclusions
The study aimed to understand a community's recovery in various inundation levels and
whether proximity to severe flooding results in a notable change; however, spatial relationships
proved more complex than initially thought. Results did not show statistically significant
differences between groups within the study area, leading to a rejection of the null hypothesis.
There are broader economic factors that overwhelmed the effects of flooding from Hurricane
Sandy, such as multiple sources of hurricane damage, economic dependence on nearby cities,
and the impact of the Great Recession.
Concerning Hurricane Sandy, this study only examined flood inundation; nevertheless,
there are more damage factors associated with hurricanes. Other aspects such as wind, heavy
rain, and storm surge could have played a role in recovery disparities. However, these factors are
not as impactful in the study area's region compared to the southern part of the United States.
Isolating damage to only one aspect lowered the statistical analysis of spatial relationships and
ignored non-coastal recovery. However, New Jersey's diversity in landscape, demographics, and
economy takes most of the responsibility in the study's resulting null hypothesis.
Broader factors contributing to the conclusion can be explained as a product of the study
area's proximity to cities and diverse elevation. Unlike other study areas examined, such as New
Orleans, New Jersey varies greatly in economic ranges and geography. The rates of urbanization,
economic growth, and socio-economic characteristics of this study area rely heavily on its
distance from the metropolis of New York and Philadelphia. Impervious surfaces of the
metropolitan areas provide an element of increased flooding that was not seen throughout the
SvIZ in the study. Instead, the clustered urban areas showed the most change due to economic
dependence on the nearby cities and lack of stormwater infiltration into the ground.

63

The recovery of the Great Recession influenced New Jersey's economic status in the time
frame of 2010 to 2018, which collapsed the housing market and affected much of the low to the
middle class. A substantial divide in income classes was created before Hurricane Sandy's
damage due to the economic crisis; therefore, the results of this study may not have been as
substantial during parallel recoveries. Nevertheless, the results did rely on sub-regionality rather
than inundation zones due to differing socio-economic composition within clusters. Based on the
hot spot analysis, trends appeared to show people moving from the state's southern region,
notably the Philadelphia metropolitan area, to the New York metropolitan area. If this trend is
accurate, the wealthiest residents moved to northern areas away from flood impact zones, while
the less affluent moved to the more impacted north and central regions.
The evaluations within the study identified disparities in income growth among the
wealthiest residents in the NIZ and SvIZ. However, Hot Spot Analysis and Spatial
Autocorrelation conveyed a more complex dynamic between census tracts. Spatial relationships
were most robust in the Vacant Housing Unit variable with a z-score of 30.05, where most of the
vacancies increased in the south and decreased in the north of New Jersey. Subregional
clustering in the state's northern, central, and southern sections suggested more dependence on
regions than impact zones. Results suggest several factors at play, and the isolated variables
examined in this study did not illuminate the entire narrative. Although the data exhibited
variation from 2010 to 2018, the results were not significant enough to claim a correlation
exclusively between economic recovery and proximity to severe flooding.

64

References
Athanasopoulou, E. 2017 "Urban coastal flood mitigation strategies for the city of Hoboken &
Jersey City, New Jersey." Doctoral Dissertation, Rutgers University.
Billings, S. B., Gallagher, E. and Ricketts, L. 2019. "Let the Rich Be Flooded: The Distribution
of Financial Aid and Distress after Hurricane Harvey." Urban Economics Association,
University of Colorado.
Bryner, N. S., Garcia-Lozano, M., and Bruch, C. 2017. "Washed Out: Policy and Practical
Considerations Affecting Return after Hurricane Katrina and Hurricane Sandy." Journal
of Asian Development 3 (1): 73-93.
Chen, V., Banerjee, D., and Lui, L. 2012. "Do People Become Better Prepared in the Aftermath
of a Natural Disaster? The Hurricane Ike Experience in Houston, Texas." Journal of
public health management and practice 18 (3): 241–249.
CityObservatory. 2018. "America’s Most Diverse, Income-Mixed Neighborhoods.”
https://cityobservatory.org/wp-content/uploads/2018/06/ADMIN_Report_18June.pdf.
Department of Homeland Security. 2016. National Disaster Recovery Framework, Second
Edition. Washington, DC. United States Government Publishing Office.
Deutz, A. 2018. “Innovative Finance for Resilient Coasts and Communities.” Accessed July 6,
2021. https://www.nature.org/en-us/what-we-do/our-insights/perspectives/-building-
coastal-resilience-through-innovation--/.
Federal Emergency Management Agency. 2017. Innovative Drought and Flood Mitigation
Projects, Final Report. Washington, DC. United States Government Publishing Office.
Felsenstein, D. and Lichter, M. 2014. “Social and Economic Vulnerability of Coastal
Communities to Sea-Level Rise and Extreme Flooding.” Natural Hazards 71 (1): 463–
491.
Finch, C., Emrich, C. T. and Cutter, S. L. 2010. “Disaster Disparities and Differential Recovery
in New Orleans.” Population and environment 31 (4): 179–202.
First Street Foundation. 2020. “The First National Flood Risk Assessment: Defining America’s
Growing Risk.” 1st Street Foundation, Inc 2020.
https://assets.firststreet.org/uploads/2020/06/first_street_foundation__first_national_flood
_risk_assessment.pdf.
Flanagan, B. E., Gregory, E. W., Hallisey, E. J., Heitgerd, J. L. and Lewis, B. 2011. “A Social
Vulnerability Index for Disaster Management.” Journal of Homeland Security and
Emergency Management 8, no. 1 (3).
Fuller, S. 2018. Using American Community Survey (ACS) Estimates and Margins of Error. US
Census Bureau, Decennial Statistical Studies Division.

65

https://www.census.gov/content/dam/Census/programs-surveys/acs/guidance/training-
presentations/20180418_MOE.pdf.
Holzer, R. 2017. “Evaluating the Minneapolis Neighborhood Revitalization Program’s Effect on
Neighborhoods.” Master’s Thesis, University of Southern California.
Howell, J., and Elliott, J. R. 2018. “As Disaster Costs Rise, So Does Inequality.” Socius 4: 1-3.
Howell, J., and Elliott, J. 2018. “Damages Done: The Longitudinal Impacts of Natural Hazards
on Wealth Inequality in the United States.” Social Problems 66 (3): 448–467.
Jacobs, J. M., Cattaneo, L. R., Sweet, W., and Mansfield, T. 2018. “Recent and Future Outlooks
for Nuisance Flooding Impacts on Roadways on the US East Coast.” Transportation
research record 2672 (2): 1–10.
Kelman, I., Gaillard, J. C., and Mercer, J. 2015. “Climate Change’s Role in Disaster Risk
Reduction’s Future: Beyond Vulnerability and Resilience.” International Journal of
Disaster Risk Science 6 (1): 21–27.
Kochhar, Rakesh. 2018. “The American middle class is stable in size but losing ground
financially to upper-income families.” Pew Research Center. Accessed November 19,
2021. https://www.pewresearch.org/fact-tank/2018/09/06/the-american-middle-class-is-
stable-in-size-but-losing-ground-financially-to-upper-income-families/.
Kopp, R.E., Broccoli, A., Horton, B., Kreeger, D., Leichenko, R., Miller, J.A., Miller, J.K.,
Orton, P., Parris, A., Robinson, D., Weaver, C.P., Campo, M., Kaplan, M., Buchanan,
M., Herb, J., Auermuller, L. and Andrews, C. 2016. Assessing New Jersey ’s Exposure to
Sea-Level Rise and Coastal Storms: Report of the New Jersey Climate Adaptation
Alliance Science and Technical Advisory Panel. Prepared for the New Jersey Climate
Adaptation Alliance, Rutgers University.
McCarthy, K., Peterson, D.J., Sastry, N., and Pollard, M. 2006. “The Repopulation of New
Orleans After Hurricane Katrina.” RAND Cooperation. Santa Monica, CA.
https://www.rand.org/content/dam/rand/pubs/technical_reports/2006/RAND_TR369.pdf.
McGhee, D. J., Binder, S. B., and Albright, E. A. 2020. “First, Do No Harm: Evaluating the
Vulnerability Reduction of Post-Disaster Home Buyout Programs.” Natural Hazards 21
(1): 05019002.
Myers, C. A., Slack, T., and Singelmann, J. 2008. “Social Vulnerability and Migration in the
Wake of Disaster: The Case of Hurricanes Katrina and Rita.” Population and
environment 29, (6): 271–291.
NJDEP. 2015. Damage Assessment Report on The Effects of Hurricane Sandy on The State of
New Jersey ’s Natural Resources. Trenton, NJ. https://www.nj.gov/dep/dsr/hurricane-
sandy-assessment.pdf.

66

Parmenter, B. M. and Lau, J. 2013. Estimating and Mapping Reliability for American
Community Survey Data. Tufts GIS Center.
http://sites.tufts.edu/gis/files/2013/11/Amercian-Community-Survey_Margin-of-error-
tutorial.pdf.
Piecuch, C. G., Huybers, P., Hay, C. C., Kemp, A. C., Little, C. M., Mitrovica, J. X., Ponte, R.
M., and Tingley, M. P. 2018. “Origin of Spatial Variation in US East Coast Sea-Level
Trends During 1900-2017.” Nature 564 (7736): 400–404.
Rasmussen, C. 2021. “Study Projects a Surge in Coastal Flooding, starting in 2030s.” Assessed
July 16, 2021. https://www.nasa.gov/feature/jpl/study-projects-a-surge-in-coastal-
flooding-starting-in-2030s.
Shao, W., Jackson, N. P., Ha, H., and Winemiller, T. 2020. “Assessing Community Vulnerability
to Floods and Hurricanes Along the US Gulf Coast.” Disasters 44 (3): 518-547.
States at Risk. 2015. “America’s Preparedness Report Card 2015: New Jersey.” Climate Central.
Assessed August 8, 2021. http://assets.statesatrisk.org/summaries/NewJersey_report.pdf.
Stocker, T.F., Qin, D., Plattner, G., Tignor, M., Allen, S.K., Boschung, J., Nauels, A., Xia, Y.,
Bex, V., and Midgley, P.M. 2013. IPCC, 2013: Climate Change 2013: The Physical
Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the
Intergovernmental Panel on Climate Change. Cambridge, UK. Cambridge University
Press.
US Census Bureau. “POVERTY STATUS IN THE PAST 12 MONTHS” The United States
Census Bureau, December 5, 2021.
https://data.census.gov/cedsci/table?t=Poverty&g=0USfalse&tid=ACSST5Y2019.S1701.
US Census Bureau. “QuickFacts: New Jersey.” The United States Census Bureau, August 8,
2021. https://www.census.gov/quickfacts/NJ.
van Holm, E. 2019. “Gentrification in the Wake of a Hurricane: New Orleans after Katrina.”
Urban studies 56 (13): 2763–2778.
Wahyuningtyas, N., Febrianti, L., and Andini, F. 2020. “The Carrying Capacity of GIS
Application for Spatial Thinking Growth in Disaster Material.” IOP conference series.
Earth and environmental science 485 (1): 1–7.
Willroth, P., Massmann, F., Wehrhahn, R. and Revilla Diez, J. 2012. “Socio-Economic
Vulnerability of Coastal Communities in Southern Thailand: The Development of
Adaptation Strategies.” Natural Hazards and Earth System Sciences 12 (8): 2647-2658.
Yabe, T., Tsubouchi, K., Fujiwara, N., Sekimoto, Y. and Ukkusuri, S. V. 2020. “Understanding
Post-Disaster Population Recovery Patterns.” Journal of the Royal Society interface 17
(163): 20190532.

67

Appendix A ACS 5-Year Estimate Attributes

Source: US Census Bureau
Evaluation Type Variable Source Table (Data Profile)
Population Evaluation Total Population Estimate
Total Population
(B01003)
Income Evaluation
Estimated Household Income Less
than $10,000
Income in The Past 12 Months
(In 2010 and 2018 Inflation-
Adjusted Dollars)
(S1901)
Estimated Household Income $10,000
to $14,999
Estimated Household Income $15,000
to $24,999
Estimated Household Income $25,000
to $34,999
Estimated Household Income $35,000
to $49,999
Estimated Household Income $50,000
to $74,999
Estimated Household Income $75,000
to $99,999
Estimated Household Income
$100,000 to $149,999
Estimated Household Income
$150,000 to $199,999
Estimated Household Income
$200,000 or more
Median Household Income (Dollars)
Median Income in The Past 12
Months (In 2010 and 2018
Inflation-Adjusted Dollars)
(S1903)
Persons Below Poverty Estimate
Poverty Status in The Past 12
Months by Sex by Age
(B17001)
Housing Evaluation
Owner-Occupied Housing Units
Households and Families
(S1101)
Renter-Occupied Housing Units
Vacant Housing Units
Vacancy Status
(B25004)

Abstract (if available)

Abstract Recent severe flooding caused by storms, such as Hurricane Sandy in 2012, has damaged vulnerable coastal communities across the United States at an increasing occurrence and severity. Not only do floods threaten lives and property, but they also alter the shape of a community through imbalanced recovery among socially and economically vulnerable populations. This concern begs the research question: what, if any, are the differences in recovery between communities of different economic standing concerning flood inundation levels after a severe coastal flooding event? Economic recovery disparity was investigated by analyzing New Jersey's socio-economic structure before and after Hurricane Sandy according to inundation depths categorized as impact zones: None (NIZ), Minor (MIZ), Serious (SrIZ), and Severe (SvIZ). The research design was developed to (1) examine the physical exposure of Hurricane Sandy across New Jersey; (2) investigate the socio-economic characteristics of New Jersey communities before and after Hurricane Sandy; and (3) determine whether, or not, proximity to severe flooding resulted in notable changes to citizen’s economic standing. The analysis compared tabular data from 2010 and 2018 American Community Survey (ACS) 5-Year Estimates using three evaluations: population, income, and housing. Results displayed variable levels of impact throughout the entire study area from 2010 to 2018 regarding population, income, and housing; however, results did not show statistically significant relationships between economic recovery and flood inundation levels.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Validating the HAZUS coastal surge model for Superstorm Sandy

PDF

Spatiotemporal analysis of the SLOSH and ADCIRC storm surge models: a case study of hurricane Ida

PDF

An analysis of racial disparity in the distribution of alcohol licenses and retailers in Orange County, California

PDF

Assessing homeless accessibility to community resources in the city of Los Angeles

PDF

Exploring the pernicious effects of redlining and discriminatory policies on an American city: a spatio-temporal case study of New York City

PDF

Impacts of vegetation management on wildfire severity: a study of the 2021 Caldor fire

PDF

Creating a flood vulnerability index for Houston, Texas

PDF

The impact of definition criteria on mapped wildland-urban interface: a case study for ten counties along the Oregon-California border

PDF

Projecting vulnerability: a combined analysis of sea-level rise, hurricane inundation, and social vulnerability in Houston-Galveston, Texas

PDF

Modeling potential impacts of tsunamis on Hilo, Hawaii: comparison of the Joint Research Centre's SCHEMA and FEMA’s HAZUS inundation scenarios

PDF

Using GIS to perform a risk assessment for air-transmitted bioterrorism within San Diego County

PDF

Geospatial analysis of the Round Fire: a replication of burn severity analyses in the Sierra Nevada

PDF

Mapping future population impacts caused by sea level rise in Huntington Beach and Newport Beach: comparing the cadastral-based dasymetric system to past dasymetric mapping methods

PDF

Investigating bus route walkability: comparative case study in Orange County, California

PDF

Assessing the value of crowdsourced data in aiding first responders: a case study of the 2013 Boston Marathon

PDF

Detection and accuracy assessment of mountain pine beetle infestations using Landsat 8 OLI and WorldView02 satellite imagery: Lake Tahoe Basin-Nevada and California

PDF

An evaluation of Esri’s tapestry segmentation product in three Southern California communities: Manhattan Beach, Santa Monica, and Venice Beach

PDF

The use of site suitability analysis to model changes in beach geomorphology due to coastal structures

PDF

Spatiotemporal visualization and analysis as a policy support tool: a case study of the economic geography of tobacco farming in the Philippines

PDF

A critical assessment of the green sea turtle central west Pacific distinct population segment utilizing maxent modeling on nesting site locations

Asset Metadata

Creator Kelly, Megan Rae (author)

Core Title The impact of severe coastal flooding on economic recovery disparities: a study of New Jersey communities following Hurricane Sandy

School College of Letters, Arts and Sciences

Degree Master of Science

Degree Program Geographic Information Science and Technology

Degree Conferral Date 2022-05

Publication Date 01/24/2022

Defense Date 01/06/2022

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

Tag economic recovery,flood,flooding,Hurricane Sandy,New Jersey,OAI-PMH Harvest,recovery disparity,relief,socio-economic

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Ruddell, Darren (committee chair), Fleming, Steven (committee member), Vos, Robert (committee member)

Creator Email MegRaeKelly@gmail.com,MRKelly@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-oUC110575731

Unique identifier UC110575731

Legacy Identifier etd-KellyMegan-10354

Document Type Thesis

Format application/pdf (imt)

Rights Kelly, Megan Rae

Type texts

Source 20220128-usctheses-batch-909 (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. The original signature page accompanying the original submission of the work to the USC Libraries is retained by the USC Libraries and a copy of it may be obtained by authorized requesters contacting the repository e-mail address given.

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