ABSTRACT 
 
 
Water Infrastructure Vulnerability to Coastal Flood Hazards: A Space-Place Analysis of 
Manteo, New Bern, and Plymouth, North Carolina  
 
by Zachary D. Oyer 
 
July 10, 2014 
 
Director: Dr. Thomas R. Allen 
Department of Geography, Planning & Environment 
 
 
With the impending threat of sea level rise, as well as the recurring annual danger 
of storm surges during hurricane season and floods from heavy rain events, North 
Carolina’s coast is especially vulnerable to coastal flooding, due mostly to large extents 
of low lying coastal areas. Water utility infrastructure is a vital resource to any 
community, while concurrently containing hazardous material that could be potentially 
devastating to the residents if parts are damaged. Unfortunately, they are commonly 
located in highly vulnerable areas along the coast. Hurricane Sandy illustrated how large 
magnitude natural hazards can damage vulnerable infrastructure, leaving municipalities 
burdened with enormous repair costs, as well as large parts of the city without running 
clean water.  
To reduce the vulnerability of these important systems in several coastal North 
Carolina communities, New Bern, Plymouth and Manteo were assessed for their 
vulnerability to storm surge, sea level rise, and riverine flooding using downscaled surge 
flood models, and applying Geographic Information Systems techniques to improve the 
accuracy of Digital Elevation Models used in flood mapping. A geospatial overlay of the 
water infrastructure assists in the computation of vulnerability of this resource to these 
risks, which will be used to promote proactive solutions to city officials in order to reduce 
their vulnerability. By modeling these different hazards for three different communities 
with different geographical contexts, we can observe how they differ within and 
throughout differing areas. 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 Water Infrastructure Vulnerability to Coastal Flood Hazards: A Space-Place 
Analysis of Manteo, New Bern, and Plymouth, North Carolina 
 
 
 
A Thesis 
Presented To 
The Faculty of the Department of Geography, Planning & Environment 
East Carolina University 
 
 
 
 
In Partial Fulfillment 
of the Requirements for the Degree 
Masters of Arts in Geography 
 
 
by 
Zachary Oyer 
July 10, 2014 
  
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
  
 
 
 
 
 
 
 
 
 
©Copyright 2014 
Zachary D. Oyer 
 
 
 
 
 
 
 
 
 
 
WATER INFRASTRUCTURE VULNERABILITY TO COASTAL FLOOD 
HAZARDS: A SPACE-PLACE ANALYSIS OF MANTEO, NEW BERN, AND 
PLYMOUTH, NORTH CAROLINA 
 
 
 
by 
Zachary D. Oyer 
 
 
APPROVED BY: 
DIRECTOR OF THESIS: __________________________________________________ 
Thomas R. Allen, PhD 
 
COMMITTEE MEMBER:__________________________________________________ 
Burrell E. Montz, PhD 
 
COMMITTEE MEMBER:__________________________________________________ 
Thomas W. Crawford, PhD 
 
CHAIR OF THE DEPARTMENT OF GOEGRAPHY, PLANNING & 
ENVIRONMENT:  
    ________________________________________________ 
Burrell E. Montz, PhD 
 
DEAN OF THE GRADUATE SCHOOL: 
 
    ________________________________________________ 
Paul J. Gemperline, PhD
ACKNOWLEDGMENTS 
 
 I would like to first and foremost thank the Department of Geography, Planning & 
Environment for allowing me to come to North Carolina and continue my education. Specifically, 
I would like to thank Dr. Tom Allen for helping me find a project and funding when I wasn’t sure 
what I wanted to do with my thesis, and helping me all along the way to finishing it. I also want to 
recognize all the help and guidance I got from Dr. Burrell Montz and Dr. Tom Crawford for being 
on my committee, and always having an open door to me when I had any questions. Also, all of 
my friends and fellow graduate students at ECU that made the whole experience much more 
enjoyable. 
 I would also like to thank all of the professors, graduate students, and friends at Texas 
A&M University who helped me lay the ground work for a great education, teaching me how to 
work harder, and get the most out of school and life. I also, of course, would like to thank my 
family who have always been around for guidance, and constant reminders for how hard work 
pays off in the end. Lastly, I would like to recognize both CISA and The North Carolina Coastal 
Atlas for providing me with the funding to make all of this possible. 
 
 
 
 
 
 
 
 
 
TABLE OF CONTENTS 
 
LIST OF FIGURES ...................................................................................................................  x  
LIST OF TABLES …………………………………………………………………………....  xii 
LIST OF ACCRONYMS …………………………………………………………………....  xiii 
CHAPTER 1: INTRODUCTION ………………………………………………………….......  1 
CHAPTER 2: RESEARCH QUESTIONS.................................................................................. 4 
CHAPTER 3: REVIEW OF LITERATURE ………………………………………………….  6 
3.1 CLIMATE CHANGE …………………………………………..............................  6 
3.2 INFRASTRUCTURE DAMAGE ………………………………………............... 10 
3.3 HEALTH RISKS ………………………………………………………………..... 12 
3.4 MITIGATION …………………………………………………………………....  13 
 3.5 GIS AND VULNERABILITY ASSESSMENT ….……………………………… 14 
3.6 COASTAL INUNDATION MODELS …………………………………….......... 15 
CHAPTER 4: METHODOLOGY …………………………………………………………....  19 
4.1 DATA ACQUISITION …………………………………………………………... 19 
4.2 DASYMETRIC MAPPING …………………….………………………………..  20 
4.3 COMPILING HIGH RESOLUTION DEMs …………………………………..... 22 
4.4 MODELING COASTAL INUNDATION ……………………………………….. 23 
4.4.1 RIVERINE ………………………………………………………….......  23 
4.4.2 STORM SURGE ………………………………………………………. 24 
4.4.3 SEA LEVEL RISE ……………………………………………………..  25 
4.5 MULTI-HAZARD MAPPING ...............................................................................  26  
4.6 VULNERABILITY ASSESSMENT ……………………………………………... 27 
4.7 HYPSOMETRIC GRAPHS …………………………………………………….... 27 
 
 
CHAPTER 5: RESULTS ………………………………………………........................…......  29 
5.1 RIVERINE ………………………………………………………………………..  34 
5.2 STORM SURGE …………………………………………………………………. 39 
5.3 SEA LEVEL RISE ……………………………………………………………..... 44 
5.4 HYPSOMETRIC GRAPHS ………………………………………....................... 49 
5.4.1 MANTEO ……………………………………………….……………...  50 
 5.4.2 NEW BERN  …………………………………………………………… 50 
5.4.3 PLYMOUTH  …………………………………………………………..  51 
5.5 MULTI-HAZARD MAPS ………………………………………………………..  52 
CHAPTER 6: DISCUSSION AND CONCLUSIONS ………………………………………  57 
6.1 DISCUSSION ……………………………………………………………………. 57 
6.2 LIMITATIONS AND IMPROVING COASTAL INUNDATION STUDIES ….. 59 
6.3 CONTRIBUTIONS OF THIS RESEARCH .......…………………………...........  61 
6.4 CONCLUSIONS ……………………………………………………………........ 63 
REFERENCES .......................................................................................................................... 66  
 
  
 
 
 
 
 
 
 
 
 
 
LIST OF FIGURES 
FIGURE 1.1: Study Areas ……………………………………………………………………....3 
FIGURE 3.1: Hurricane Irene High Water Marks ……………………………………………... 9 
FIGURE 3.2: Tail Water …………………………………………………………………….... 11 
FIGURE 4.1: Dasymetric Mapping ...........................................................................................21 
FIGURE 4.2: Flow Diagram ..................................................................................................... 23 
FIGURE 4.3: Downscale Methodology ..................................................................................... 25 
FIGURE 4.4: Hypsometric Graph Example .............................................................................. 28 
FIGURE 5.1: Infrastructure of Manteo ..................................................................................... 31 
FIGURE 5.2: Infrastructure of New Bern ................................................................................. 32 
FIGURE 5.3: Infrastructure of Plymouth .................................................................................. 33 
FIGURE 5.4: Riverine Inundation for Manteo ......................................................................... 36 
FIGURE 5.5: Riverine Inundation for New Bern ..................................................................... 37 
FIGURE 5.6: Riverine Inundation for Plymouth ...................................................................... 38 
FIGURE 5.7: Storm Surge Inundation for Manteo ................................................................... 41 
FIGURE 5.8: Storm Surge Inundation for New Bern ............................................................... 42 
FIGURE 5.9: Storm Surge Inundation for Plymouth ............................................................... 43 
FIGURE 5.10: Sea Level Rise Inundation for Manteo ............................................................. 45 
FIGURE 5.11: Sea Level Rise Inundation for New Bern ......................................................... 46 
FIGURE 5.12: Sea Level Rise Inundation for Plymouth .........................................................  47 
FIGURE 5.13: Vulnerability Summary .................................................................................... 48 
FIGURE 5.14: Manteo Hypsometric Graph .............................................................................  49 
FIGURE 5.15: New Bern Hypsometric Graph .........................................................................  50 
FIGURE 5.16: Plymouth Hypsometric Graph .......................................................................... 51 
FIGURE 5.17: Annotated Multi-Hazard Map ........................................................................... 53 
 
 
FIGURE 5.18: Manteo Multi-Hazard Map .............................................................................. 54 
FIGURE 5.19: New Bern Multi-Hazard Map ..........................................................................  55 
FIGURE 5.20: Plymouth Multi-Hazard Map ...........................................................................  56 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
LIST OF TABLES  
TABLE 4.1: Data Sources ......................................................................................................... 20 
TABLE 5.1: Infrastructure and Population ............................................................................... 30 
 
  
 
 
LIST OF ACRONYMS 
CRC – Coastal Resource Commission 
DEM – Digital Elevation Model 
DFIRM – Digital Flood Insurance Rate Map 
GIS – Geographic Information Systems 
HECRAS – Hydrologic Engineering Centers River Analysis System   
IPCC – Intergovernmental Panel for Climate Change 
LiDAR- Light Detection and Ranging 
MOM – Maximum of Maximums 
NCFMP – North Carolina Floodplain Mapping Program 
NC-SLRMS – North Carolina Sea Level Rise Management Study 
NHD – National Hydrography Dataset 
NOAA – National Oceanic and Atmospheric Association 
NWS – National Weather Service 
SLAMM – Sea Level Affecting Marshes Model 
SLOSH – Sea, Lake, and Overland Surges from Hurricanes 
USGS – United States Geological Survey 
 
  
 
 
CHAPTER 1: INTRODUCTION 
A variety of hazards threaten coastal communities in North Carolina, including sea level 
rise, storm surge, coastal riverine flooding, and shoreline erosion. With about 40% of the United 
States population on the coast, these hazards put many people’s lives at risk, as well as many 
residential and commercial buildings, and critical urban infrastructure (U.S. Census Bureau, 
2011). In order to help reduce risk and future disaster damage owing to these natural hazards, 
coastal communities are wise to assess their vulnerability and evaluate where they are at most risk 
of damage. With increasing availability and accuracy of Light Detection and Ranging (LiDAR) 
and associated geospatial technology, particularly accurate models can be created to simulate 
inundation arising from the different processes of coastal flooding. With these compounding 
effects influencing vulnerability to coastal flooding, some models predict an increase of up to 30 
million people affected by this threat every year worldwide by the year 2080 (Nicholls et al., 
1999).  Other modeling by Small and Nicholls (2003), predicts an increase in people living within 
a 100 year storm surge floodplain, from 200 million in 1990, to up to 600 to 800 million by 2100.  
 Water utility infrastructure, which involves the treatment and transportation of both clean 
water and sewage, is often located in low-lying areas along the coast, making these a highly 
vulnerable component of local utilities. A working definition of water utilities includes material 
infrastructure, or a public facility, which is physically comprised of canals, waterways, 
waterworks, reservoirs and pipelines for the purpose of supplying drinking water, industrial-use 
water, irrigation water, and wastewater disposal functions (Buhr, 2003).  Flooding can cause 
pumping problems as the inundated water alters the hydraulic head of the system, as well as 
causing debris blockages and pipe failures (Titus et al., 1987, Chughtai and Zayed, 2008). 
Additional concerns for water infrastructure damage include the intrusion of salt water, which 
2 
 
may lead to pipe and valve corrosion (Fugro, 2012). Reduction or elimination of human and 
environmental exposure to dangerous materials, such as sewage, is very beneficial to at risk 
communities. Several studies have assessed the vulnerability of such utilities to storm surge or sea 
level rise in coastal communities and determined this infrastructure to be highly vulnerable 
(Burkett and Davidson, 2012, Heberger et al., 2011). Burkett and Davidson concluded that clean 
water and sewage treatment utilities were at significant threat from the combination of sea level 
rise and storm surge in the United States.  
 The vast destruction Superstorm Sandy caused in the Mid-Atlantic in 2012 highlights our 
coastal vulnerability. Estimates for the cost of reconstruction from the damage caused by Sandy 
ranges between $140 to $240 billion (Bloomberg, 2012). With such critical and expensive 
infrastructure in highly vulnerable areas of coastal cities, expectations are for this trend to 
continue to increase if mitigation or adaptation measures are not undertaken to reduce our 
vulnerability and avoid ensuing damage. In the wake of Sandy’s destruction, New York’s 
Governor Andrew Cuomo vowed to rebuild and strengthen several aspects of water infrastructure 
to decrease its vulnerability to coastal hazards again (NY.gov 2100 Commission, 2013). 
Strategies that balance the investments in mitigating damages to existing infrastructure and seek 
to develop more adaptive, resilient infrastructure in the future are gaining serious consideration.   
In order to contribute to our understanding of adaptation, mitigation and preparation for 
these increasing coastal hazards, this study evaluates vulnerability of water utility infrastructure to 
coastal flooding of North Carolina cities and towns, including the towns of Manteo and Plymouth 
and the City of New Bern (Figure 1.1). This study adopts a comparative approach to assess how 
local coastal geography can influence the risks posed by different hazards over space and time. 
We have the ability to not only assess how vulnerable the infrastructure is currently, but also to 
3 
 
speculate how that may change over time as climate change and coastal processes interact with 
sea-level rise. In addition, the investigation can reveal differential spatial and time-evolving risks 
for communities, allowing them to prepare and plan for change into the future. Lastly, and most 
importantly, the goal is to approach hazards differently, by analyzing three separate flood hazards 
in three communities with different geographical settings. This type of approach will provide 
more context to true flood vulnerability that many other studies fail to assess.  It is hoped that this 
study will encourage coastal communities to be proactive in seeking solutions to the ongoing and 
potentially increasing risk of flood hazards. 
 
Figure 1.1: Study area and communities of interest in eastern North Carolina. 
 
 
CHAPTER 2: RESEARCH QUESTIONS 
1. How do riverine flooding, storm surge, and sea level rise risks threaten the coastal 
communities of Manteo, New Bern, and Plymouth, North Carolina?  
While different geographic regions experience different risks and hazards, these hazards all 
contribute to the risk of all three communities.  Vulnerability to hazards such as sea level rise and 
storm surge is increasing in coastal regions (Hecht, 2006). With North Carolina’s low-lying 
coastal plain, storm surges are a continual threat, while the rise in sea level is a chronic, long-term 
threat facing the entire state’s coastal region. Answering this question will elucidate the relative 
current threats of these hazards and allow for the comparison among the three locations being 
evaluated: 1) a river-dominated community (Plymouth), 2) a hybrid riverine-estuarine setting 
(New Bern) and 3) a coastal estuarine island community (Manteo).  
2. How vulnerable is the water infrastructure of Manteo, New Bern, and Plymouth to coastal 
flood hazards? 
While proactive planning efforts should keep important infrastructure out of flood plains, and 
away from harmful coastal hazards, wastewater treatment plants are often placed in highly 
vulnerable areas. The storm surges from larger storms will most likely encompass much of the 
infrastructure currently in vulnerable areas. With a continued rise of sea level, likely at an 
accelerated rate (Vermeer and Rahmstorf , 2009), there is an added long term flood hazard risk to 
this critical infrastructure, even if a large enough storm surge is not generated in the near future. 
As sea level rises, marshes and higher tidal zones begin to impact storm and wastewater outfalls, 
increasingly threatening the infrastructure with tailwater flooding and eventually outright flooding 
of facilities such as pump stations and underground pipes. Marshes may even invade facilities and 
reclaim transportation corridors. It is expected that all of these communities will have at least 
some vulnerability to all of these hazards, which can always be reduced. 
5 
 
3. Which coastal hazard poses the greatest risk in the future for each of these communities’ 
water resources infrastructure? How do these threats to infrastructure vary spatially and 
temporally among the study sites looking into the future?  
A cursory inspection of the spatial placement of the infrastructure, revelas high 
vulnerability to coastal floods. Larger storm surges should cause the most flooding and affect the 
most infrastructure since it has the quickest and largest influx of water. Regional sites should be 
differentially vulnerable to other hazards, depending on the geography of the coast.  Greater 
exposure to open bodies of water will create more vulnerability to storm surges, while larger river 
basins will create more riverine vulnerability. By discovering which hazards are more of a threat 
to each community, this information could be useful moving forward with adaptation and 
mitigation efforts. 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
CHAPTER 3:  REIVEW OF LITERATURE 
Climate change and sea level rise have become topics of great interest in recent years. 
With almost 40% of the U.S. population living in coastal counties, which make up only 10% of 
the total land cover, these counties have a population density of  446 people per square mile (171 
people per square km), versus  the national average of 105 people (40 per square km) for the 
contiguous United States (U.S. Census Bureau, 2011). In North Carolina, Dare County comprises 
a large portion of the Outer Banks barrier island chain and was the fastest growing North Carolina 
county between 1970 and 1995 (Overton et al., 1999). With such a high concentration of the 
population along the coast, and a high value of coastal property, some estimates predict that a 
large coastal storm could easily exceed $1 billion in damages on the North Carolina coast alone 
(Hondula and Dolan, 2010). The future risks such coastal communities face all over the world 
arise not only from coastal flooding from storms but from continued sea level rise as well. 
3.1 Climate Change 
Sea level rise is measured by two means, a global measurement of worldwide sea level, or 
“Eustatic” sea level, and a more local measurement termed “Relative Sea Level”. The eustatic sea 
level rise rates have been measured at about 3.4 mm per year through altimetric means, with about 
30% of that rise accounting for thermal expansion, and about 55% accounting for land ice melt 
(Cazenave and Llovel, 2010). The measured relative sea level rise of North Carolina has varied 
from 3.6 to 4.5 mm per year (Kemp et al., 2009a, Kemp et al., 2009b), exacerbated by strong 
local subsidence, particularly in the northeastern coastal plain and Outer Banks. Predictions for 
the rise in eustatic sea level have been generated in many reports, with many differing opinions on 
the rate of change. The Intergovernmental Panel on Climate Change (IPCC), which is recognized 
as one of the leading international organizations on climate change and sea level rise, predicts that 
7 
 
by the year 2100, sea level will follow a linear rate of change, and will experience between about 
20 to 60 cm of rise (IPCC, 2007). However, many studies suggest that the rate of sea level change 
will be an exponential rise. Vermeer and Rahmstorf (2009) point out that the IPCC report lacks an 
inclusion of dynamic ice processes, and predict that by the year 2100, sea level could be 1 to 1.5 
m higher than it currently is. Concurrently, as sea level rises, not only will it continue to inundate 
more land, but it will also change the shape of the coast over time (Davidson-Arnott, 2005). With 
the rise of sea level expected to occur for a long period of time, there is much need for these 
coastal communities to assess the risk of coastal land loss, and to make preparations, mitigate or 
adapt.  
 In the short term, it is simplistically observed that more coastal inundation occurs annually 
from storm surges than annual sea level rise. North Carolina experiences strong  thunderstorms, 
tropical cyclones, and nor’easters, with nor’easters creating the most wave energy on North 
Carolina’s coast each year, bringing seven times the amount of total storm wave energy produced 
by hurricanes (Smith et al., 2006). Storm surge flooding, coupled with heavy rainfall from these 
storms, causes extensive flooding and massive damage to the entire developed coast. One of the 
most famous nor’easters in North Carolina history, the Ash Wednesday Storm of 1962, impacted 
over 1,000 km of the Atlantic coast shoreline, causing over $300 million in property damage with 
over 10 m high waves (Davis and Dolan, 1993). 
 Tropical cyclones can be quite destructive as well, with the 2005 hurricane season being 
the most costly in recorded history. In that season, hurricanes were the cause of nearly 1700 
deaths, and caused well over $100 billion in damages in the United States (Beven et al., 2007).  
Hurricane Floyd brought some of the worst flooding in North Carolina’s history in 1999, brought 
upon by torrential rainfall rather than extreme storm surges. Compounded by a wet summer and 
 8 
 
two preceding tropical storms, the flooding associated with Hurricane Floyd was categorized as a 
500 year flood event, and over the next month drained a volume of water almost equal to the 
entire amount of the Pamlico Sound (Bales, 2003). Unlike many hurricanes associated with high 
death tolls and extensive flood damage, most of the flooding from Floyd was riverine, and 48 of 
the 56 deaths from Floyd were due to freshwater flooding (NWS, 2013). The transport of flood 
waters from Floyd into the Albemarle and Pamlico Sounds caused severe flooding for towns near 
river embayments and estuaries such as New Bern (Neuse River) and Plymouth (Roanoke River), 
but had comparatively very little effect on the Outer Banks at Manteo. This flood disaster 
highlights the spatial differences of some coastal flood hazards and stresses the importance of a 
comparative, multi-hazard approach.  
Another recent storm that informs our understanding of the differences between flooding 
among these study sites is Hurricane Irene, which caused extensive damage in North Carolina as 
well. On August 20, 2011, Hurricane Irene was the first hurricane to hit the United States since 
Hurricane Ike hit the Texas coast in 2008, and was the 10th billion dollar disaster in the United 
States, causing 300,000 people to evacuate in North Carolina alone (NOAA, 2011). The United 
States Geological Survey (USGS) has a database of high water marks (USGS, 2014), which were 
extracted to compare the three study locations, and the flooding that occurred differently among 
them (Figure 2.1). New Bern recorded the highest flood heights at over 2.2 m, while Manteo’s 
highest recorded flood elevation was 2.1 m, and Plymouth only recorded a flood elevation of 1.98 
m several kilometers down the Roanoke River. These differences illustrate the extreme variability 
of event impacts across a complex, extensive coastal plain and estuarine system such as eastern 
North Carolina.  
9 
 
 
Figure 3.1: Heights of high water marks (meters) from Hurricane Irene in 2011 for A) New Bern 
B) Plymouth and C) Manteo overlaid onto orthophotography (composited of 2010, 2011 and 2013 
imagery). 
 
With climate change, the behavior of tropical cyclones is expected to change (IPCC, 
2007). High resolution modeling is improving the ability to predict these changes of storms with 
warming temperatures and sea level rise. Most importantly, these models predict an increase in 
intensity of 2 to 11% by 2100, but yet also a decrease in total hurricane frequency by 6 to 34%. 
However, there may be an increase in the number of major hurricanes (Knutson et al., 2010). An 
10 
 
increase in hurricane intensity brings stronger winds, which will push more water onto shore at 
landfall and create higher storm surges. The combination of more intense hurricanes with larger 
storm surges, in addition to accelerated sea level rise makes predicting the extent of coastal 
flooding more difficult and could possibly result in under-estimations in the predictions.  
3.2 Infrastructure Damage 
For water infrastructure, coastal flooding causes many problems with pumping pressure 
and can also cause extensive damage to sewer pipelines from debris blockage (Titus et al., 1987). 
Pumping pressure is determined by the “hydraulic head”, which is the measured liquid pressure 
above a datum, which constitutes the water treatment plant’s ability to pump water to its 
beneficiaries. The higher the water table rises above this datum, the more hydraulic pressure 
lessens, and the plant can lose the ability to pump water.  
The integrity of the pipes may also be compromised due to coastal hydrologic processes 
such as groundwater infiltration (Chughtai and Zayed, 2008). Issues stemming from groundwater 
infiltration occur mostly with the soil surrounding the pipelines. The flowing of the groundwater 
around the pipes may erode the soils surrounding the pipes and undermine their support. With 
inadequate support from the ground, the pipes can come under too much stress, resulting in pipe 
failure. Additional damage to underground infrastructure can be incurred through the migration of 
saline water from tail water. The tail water is simply the elevation of the water downslope from a 
structure, such as a dam, or a coastal outfall for run off or treated wastewater (Figure 2.2). Once 
tail water elevations are high enough to reach the entrance of a coastal outfall, the slope of the 
water level landwards will push the direction of flow up the outfall, which can cause corrosion to 
pipes, valves and fittings (Fugro, 2012). The salinity of the water will also have an effect on the 
11 
 
magnitude of corrosion, with rising water salinity causing corrosion to occur faster than before 
(Fink, 1960).   
 
Figure 3.2: Diagram of raising tailwater elevations migrating upland and up coastal outfalls. 
 
These processes involving saline groundwater will increase the risks for Manteo as 
compared to New Bern or Plymouth, owing to its location closer to the open ocean, and to its 
higher salinity surrounding estuaries by proximity to Oregon Inlet. The sounds of North Carolina 
contain brackish water, which can have a range of salinities as fresh water from tributaries meets 
the saline waters of the ocean. The salinity of these waters varies based on several processes, 
particularly proximity to fresh or salt water sources. However, as sea level rise increases, the 
higher salinity of the ocean will intrude into the estuaries, raising their salinity, and increasing 
their risk to corrosion as well. Some studies predict an increase of salinity of up to 4 ppt in 
drinking water with 100 cm of sea level rise in the Chesapeake Bay tributary rivers, far surpassing 
the EPA standards for dissolved solids in drinking water, which is 0.5 ppt (Rice et al., 2012).  
12 
 
Preparation and mitigation for coastal communities can help reduce these costly repairs and 
environmental damage due to failing utility infrastructure. 
3.3 Health Risks 
 In addition to the potential risk of increasingly saline water on infrastructure, flooding and 
relative sea-level rise also pose acute health hazards. With damage and possible failure of water 
infrastructure comes the possible spillage of dangerous material such as raw sewage, leading to 
associated health risks. There is no question that raw sewage in the public water supply is a health 
risk to the entire community (Lane et al., 2013), however, there are serious negative effects to the 
entire ecosystem as well. The effluent release of this fecal matter floods the system with nutrients 
in coastal environments, causing massive algal blooms which create severe hypoxic conditions 
and substantial fish kills. While these contaminants may decline in the water column at an 
exponential rate through bacterial consumption, their presence in the sediment can last much 
longer, leaving the area contaminated for an extended period of time (Mallin et al., 2007). 
Downstream ecosystem impacts can affect fisheries and livability and amenity value of coastal 
recreation and tourism. Fish kills, algal blooms, and reduced fishing are just a sampling of 
possible additional ecological impacts that could degrade the coastal economy. 
 Additional risks from flooding to humans, besides drinking water contamination from 
sewage, are direct exposure to sewage overflow, living in flood damaged homes containing mold 
or with utility outages for an extended period of time, and even drowning during flooding events 
(Lane et al., 2013).  These impacts to the well-being and quality of life of citizens should be 
among the top priority reasons for assessing risks and reducing the vulnerability of water 
 13 
 
infrastructure. While cost-benefit analysis should be considered for mitigation efforts, the effect 
on human lives and ecosystem goods and services should also be assessed.  
3.4 Mitigation 
As sea level rise becomes a more recognized threat to coastal communities, more cities 
have examined options to be more prepared to deal with these impending risks, in the form of 
policy and physical preparation (Munaretto et al., 2012). The ability to act proactively to the 
anticipation of coastal flooding requires a combination of wise development and conservation of 
beaches and marshes along with considering potential hardened structures like levees, dams and 
bulkheads. Highly vulnerable and important infrastructure, such as sewage plants and pipelines 
should be moved to less vulnerable areas to prevent sewage backup and spills to save money in 
repairing damage.  To accomplish these important and often expensive tasks, political cooperation 
and foresight are very important (Munaretto et al., 2012).   
 To assess the need for proactive solutions to coastal inundation, many cities have 
evaluated their own vulnerability to future sea level rise and storm surges (Heberger et al., 2011, 
Friedrich and Kretzinger, 2012). Assessments like these using GIS to model coastal inundation 
using high resolution digital elevation models (DEMs) with geospatial analysis of geographic 
locations of important infrastructure can provide important information to cities. While some 
studies focus more on vulnerability to sea level rise (Friedrich and Kretzinger, 2012), others 
combine risk assessment of floodplain inundation and surges with local estimates of sea level rise 
(Heberger, 2011). With information on thse hazards overlain, city planners and local government 
officials can consider how to move forward, improving the resilience of existing infrastructure, 
and designing for future projects with lower risks.  
14 
 
Resilience is the ability of a system to bounce back to its initial state after a disaster, such 
as flooding, which is a term adopted from physics, describing a material’s ability to return to 
homeostasis after a disturbance (Norris et al., 2008). To improve resilience, communities can 
mitigate and adapt through several means.  Mitigation and adaptation are both strategies to reduce 
a community’s vulnerability to risk, but they differ by their goals in implementation. Kundzewicz 
and Matczak (2012) describe this difference as such: mitigation treats the causes of problems, 
while adaptation seeks to treat the symptoms. That is, mitigation aims to decrease the threat that 
causes coastal flooding, climate change, through measures such as reduction in carbon emissions, 
while adaptation attempts to decrease the damage that flooding causes, such as moving or 
elevating buildings.  
3.5 GIS and Vulnerability Assessment 
 In order to create accurate assessments of flood inundation, it is necessary to correct 
digital elevation models used in geographic information systems (GIS) to portray the terrain in a 
more accurate manner than the publicly shared DEM. Hydro-correction is the process of editing 
the modeled terrain in order to adjust the hydrology of the system for more accurate flood models 
than before the correction. The use of hydro-correction can change the results of flooded areas 
drastically, in some studies changing the area of inundated regions by 760 km2 (Poulter and 
Halpin, 2008). The change in the DEM creates flooding in areas that may be misrepresented by 
conventional LiDAR data, such as ditches, that may not have been captured by the models before 
the hydro-correction. Other forms of hydro-correction can be the addition of culverts, which along 
with ditches, can greatly affect the flow of water, hydrography, and delineation of watershed size 
and shape (Duke et al., 2003).  
 
15 
 
To assess of vulnerability in this study, we must first define explain a working definition. 
Several studies have made attempts to create unified definitions of this term in order to help bring 
the scientific community into agreement on how to use and measure it. Wolf, et al. (2013), have 
defined vulnerability assessments dealing with climate change as the measured degree of effect a 
stimulus has on an entity. The IPCC has created a definition as well, which is “the degree to 
which a system is affected, either adversely or beneficially, by climate variability or change” 
(Parry et al., 2007). Vulnerability can be calculated and represented through different 
measurements, such as a relative or absolute count. Relative vulnerability measures vulnerability 
based on a percentage of the total system affected, while an absolute vulnerability is based on a 
raw count of affected items. Relative vulnerability will tend to have higher values for 
communities with less infrastructure, such as towns with only one wastewater treatment plant. 
Meanwhile, absolute vulnerability assists in conveying vulnerability for larger communities with 
more infrastructure. For example, if you compare a town with one treatment plant with a city with 
four treatment plants, and both have one at risk treatment plant each, their respective relative 
vulnerabilities will be 100% and 25%, while both have an absolute vulnerability of 1. Relative 
vulnerability allows for comparison between communities of different sizes, while an absolute 
count will create bias on size, but will also give a different perspective on vulnerability. 
3.6 Coastal Inundation Models 
 The Sea, Lakes and Overland Surge from Hurricanes (SLOSH) model was created by the 
National Weather Service (NWS) of the National Oceanic and Atmospheric Administration 
(NOAA) to provide a fast, spatially universal, and relatively precise method to predict storm 
surge extents for oncoming storms (Zhang et al., 2008). The model uses size, position and 
intensity of the hurricane to calculate the hypothetical inundation extent, excluding measurements 
16 
 
such as wind parameters and tides (Jelesnianski et al., 1992). It is a two dimensional model that 
uses a polar grid system and differing cell sizes to derive a predicted storm surge height with an 
accuracy of about ± 20% when tested against field-measured heights (Zhang et al., 2008). The 
20% rule is a conservative estimate, reflecting the NWS concerns for cascading errors from storm 
track and intensity variation as well as the fact that SLOSH does not explicitly include either 
astronomical tide, waves or wave run up. However, the SLOSH model has predicted with far 
greater accuracy than the standard 20% rule in some instances. For Hurricane Katrina, the 
predicted storm extent matched high water marks at ± 5% (Zachry et al., 2012). SLOSH is used 
conservatively, usually with pre-run libraries of models for various storm tracks and intensities, 
and applicable tide and wave heights are added to the forecast and briefings from other model and 
observation sources. The SLOSH model uses simple meteorological parameters to compute the 
forces that drive the water onto land, using generalized and universal coefficients for variables 
such as surface friction to help make the model easier to use (Jelesnianski et al., 1992). In 
addition, SLOSH remains the current operational model for the NWS, and it will allow for this 
study to evaluate the different hurricane categories to generate flood risk areas for each one.  
 There are several methods scientists use to project how raising sea levels will affect the 
coast, one of which is the Sea Level Affecting Marshes Model (SLAMM), which will help to 
assess the wetland response to inundation from sea level rise. The SLAMM model was created by 
Park et al. (1986) to simulate how coastal land cover will be affected by sea level, and has 
assisted in environmental decision making since its creation. This model uses sea level rise 
scenarios taken from the IPCC, and uses not only topography to compute inundation, but can also 
estimate changes in land cover, and relative changes in erosion based on the coastal geometry 
17 
 
(Clough et al., 2009). The six processes it includes to determine how wetlands are affected are 
inundation, erosion, overwash, saturation, accretion, and salinity (Clough et al., 2009).  
 Another, more simplistic method of modeling coastal inundation from sea level rise is 
with a simple “bathtub” model. This type of model involves a manual raising of the sea level by 
the modeler. In a typical bathtub model, hydro-connectivity to a water source, such as the ocean, 
is not accounted for, and cells are flooded solely on the basis of the being above or below the 
desired flood elevation. However, in order to create accurate sea level rise scenarios, studies have 
modified this approach to account for hydro-connectivity using the cost distance tool in ArcGIS 
(Allen et al., 2013). As opposed to the SLAMM models, these models predict only direct 
inundation from sea level rise, rather than the effects of land cover type migration. Also, the 
SLAMM model is limited to the resolution of satellite land cover data, which is delineated at a 
coarse 30 m resolution. The bathtub scenarios, however, can be scaled to the size of the DEM, 
which in the case of this study, is a much finer scale than the land cover resolution.  
 Flooding by rainfall run off, or riverine flooding, is represented through FEMA Digital 
Flood Insurance Rate Maps (DFIRMs). The State of North Carolina conducted a project in 2001 
called the North Carolina Floodplain Mapping Project to create a digital database of the existing 
FIRMs of both the 100 and 500 year floodplains. These floodplains are created using water 
inputs, discharge and topography to calculate areas of inundation. In one study, DFIRMs were 
tested against field-measured floodplain boundaries, and the 100 year floodplain DFIRMS 
captured about 67% of the true flooded parcels (Aycock and Wang, 2004). Another option for 
riverine flood models would be to create our own extents using the HEC-RAS model. This model 
uses one-dimensional energy equations using cross sections along a basin to calculate flooding, 
accounting for flow rates and friction with the bottom (Kelly et al., 2007). However, the DFIRMs 
18 
 
from the NCFMP represent the same type of models included in the HEC-RAS or other similar 
models, and will represent the extent of the floodplains in this study. Periodically, these DFIRMs 
are updated as new and higher resolution LiDAR data becomes more available.  
  
 CHAPTER 4: METHODOLOGY 
 This project was accomplished almost entirely with GIS, through which most of the 
spatial analysis, and modeling will be undertaken. The study will be divided into four main 
sections: 1) data acquisition, 2) compilation of high-resolution DEMs, 3) modeling of sea-level 
rise, flooding and storm surge, and 4) analysis and synthesis of results, including graphics and 
dissemination with community officials and planners for possible incorporation in mitigation and 
adaptation plans.  The driving climate change scenario for future coastal evolution from sea-level 
rise used is the local relative sea-level change in the NC Sea-Level Risk Management Study (NC-
SLRMS) as proposed in the 2010 NC Science Panel Task Force of the NC Coastal Resources 
Commission (NCCRC Science Panel 2010).   
4.1 Data Acquisition 
 As with many GIS projects, acquiring all the data from different sources to pull together 
into one comprehensive map is not always an easy task. Several different sources were explored 
for GIS layers to represent utility water infrastructure such as water and sewage pipelines, water 
and sewage treatment plants, as well as pumping stations and service areas. Other data needed for 
this project are ditches and the high resolution (6.1 m or finer) DEMs for the three communities 
of New Bern, Manteo and Plymouth. 
 High resolution DEM data were acquired in prior work of the East Carolina University 
(ECU) Renaissance Computing Institute (RENCI) engagement center from NC Floodplain 
Mapping. Data acquired through other sources include sewer service areas, pipelines, pumping 
station, treatment plants, and ditches (Table 4.1). The ditches were found through the National 
Hydrography Dataset (NHD) provided by the USGS. The plants, pipelines, and service areas are 
accessible by the North Carolina One Map, an online dataset that provides GIS data to the public. 
20 
 
After contacting local professionals, more accurate data for New Bern than was previously found 
on OneMap, such as clean and wastewater lines, and pumping and booster stations were found 
through a published webmap provided by Kevin Gaskins (GIS technician, City of New Bern). 
Table 4.1: Sources and publication dates for layers needed for vulnerability assessment. 
 
4.2 Dasymetric Mapping 
 Using the service areas of the wastewater treatment plants, it is necessary to estimate the 
number of people who would be affected by inundated treatment plants that could no longer serve 
the public. In order to derive this population count, the population was distributed spatially 
through dasymetric mapping. To do this, census block data from 2012, along with land coverage 
satellite data that has been reclassified into “high intensity”, “low intensity”, “cleared land” and 
“uninhabited” was used. The dasymetric mapping tool, provided publicly by the USGS, 
redistributes population totals for the census block groups based on the land coverage, and gives a 
population count in a pixel matching the resolution of the land cover data, in this case, 30 m. The 
totals are be placed into the categories of high, low and cleared land, while the uninhabited areas, 
such as water or swamps, get a population count of 0 (Figure 4.1). Using this, a zonal statistic tool 
21 
 
was run to find the sum of all the pixels within each service area, and find the population affected 
by inundation.  
 
Figure 4.1: Example of dasymetric population mapping to show population redistribution for a) 
New Bern, b) Manteo, c) Plymouth. Population data provided by the 2010 Census, land cover 
data provided by the 2010 NOAA Coastal Change Analysis Program, and orthoimagery 
composited by NC OneMap with 2010, 2011, and 2013 imagery. 
 
22 
 
4.3 Compiling and Improving High-Resolution DEMs  
Part of what makes this project distinctive is the accuracy of the DEM for water flow. 
While the resolution of the DEMs is high, important hydrologic features may still be excluded 
from floodplain LiDAR that may affect the direction of water flow drastically for different 
hazards. In order to solve this problem, the project aimed to improve the DEMs by additional 
hydro-correction to “burn” the ditches into the DEM to create localized areas of higher resolution 
hydrography than before. The ditch lines brought in from the NHD were overlaid on the DEM to 
show where the DEM needed to be updated, while additional areas, verified through 
orthophotographs as other ditches, were manually digitized. To burn the streams into the raster, 
the minimum elevation value in the ditch line is applied to all the remaining cells within that line. 
A conditional statement is then used to incorporate the ditch elevations where they are present, 
and the DEM values where they are not (Figure 4.2). With the flooding of ditches, the 
hydrography adjusts and lessens the flooding in other areas with higher elevation that are less 
likely to inundate in reality.  
23 
 
 
Figure 4.2: Flowchart describing the process of hydro-correcting the DEM. Blue ovals are data 
inputs, the orange box is a statistical calculation taken outside of ArcGIS, the red boxes are tools 
ran in ArcGIS, and the green pentagon is the output. The dotted line from DEM to the “line to 
raster” tool represents the input of the DEM’s cell size for the rasterized ditch lines. 
 
4.4 Modeling Coastal Inundation 
 This study incorporates three types of coastal flooding for these coastal communities: 1) 
riverine flooding, 2) storm surge, and 3) sea level rise. Each flood risk model is represented 
through different means. The riverine floodplains are models already generated through FEMA, 
while the sea level rise and storm surge downscale models are generated in this study. 
4.4.1 Riverine model 
The riverine floodplains are shown through DFIRMs created by FEMA, and show both 
the 100 and 500 year floodplains. To represent these floodplains in this study, they were acquired 
through the NCFMP and brought into ArcGIS to overlay the water utility infrastructure for 
geospatial analysis. The 100 and 500 year floodplain layers were extracted into two separate data 
24 
 
layers, and were visualized through the definition query tool to represent the corresponding 
floodplain for vulnerability assessment.  
4.4.2 Storm Surge model 
The storm surge floodplains were generated using the SLOSH Maximum of Maximum 
(MOM) outputs for slow moving hurricanes in the Cape Hatteras/ Pamlico Sound basin. The 
MOM outputs give us worst-case hurricane inundation scenarios for all five categories of storms, 
excluding wave and tide elevations. However, the scale of these outputs for local communities is 
not precise enough to give real insight into their actual vulnerability. To adjust for this, we 
downscale the resolution of the SLOSH output grid to the resolution of the local DEM (Figure 
4.3). After the resolution of the SLOSH is matched to the DEM, we then deem each pixel to be 
“floodable” or “non-floodable” based on the difference of the inundation calculation and the 
elevation of the DEM. Then, in order to generate more accurate flood zones, we must also 
account for hydro-connectivity from the ocean. To achieve this, a cost-distance function is used in 
the model to make unconnected raster cells too “expensive” to include (Allen et al., 2013). These 
maps give a more direct map of actual inundated cells rather than a proxy of land cover changes 
from the SLAMM model for sea level rise scenarios.  
25 
 
 
Figure 4.3: Methodology of downscaling SLOSH outputs to local DEM resolution at New Bern, 
NC. A) SLOSH output polygons with extent indicator for view shown in B, B) Inset in A, 
showing relative pixel resolution to SLOSH grid, C) SLOSH grid is converted to raster and 
matched with DEM pixel resolution, so that each DEM pixel has a corresponding SLOSH surge 
height calculation. The raster calculator can then calculate the difference between elevation height 
and surge height to find “floodable” (red) and “non-floodable” (yellow) pixels, D) After hydro-
connectivity is simulated using the cost-distance tool, isolated flooded regions are excluded that 
would not flood in reality. 
4.4.3 Sea Level Rise 
 Sea level rise is shown through a manually raised “bathtub” type scenario in increments 
of 40 cm, 60 cm, 80 cm, 100 cm, and 150 cm. This range of scenarios is derived from the North 
Carolina Coastal Resource Commission (CRC) Science Panel on sea level rise. The CRC is 
composed of North Carolina researchers who consulted several papers mentioned above, such as 
Kemp (2009), Vermeer and Rahmstorf (2009), and others, to generate sea level rise predictions 
26 
 
for the state of North Carolina, rather than global estimations from the IPCC. The methodology 
for the sea level rise model is a similar, yet simpler version to the storm surge model described 
above. However, instead of using pre-run modeled outputs such as SLOSH, DEM pixels were 
simply reclassified as “floodable” or “non-floodable” based on being greater or less than the flood 
scenario ran, e.g. 40 cm, 60 cm, 80 cm, 100 cm, or 150 cm. Then, the same cost distance tool was 
used to simulate hydro-connectivity as was used in the storm surge downscale model.  
4.5 Multi-Hazard Mapping 
There are several methods available to potentially visualize inundation and vulnerability. 
Most simply, the infrastructure is mapped with each hazard type with all scenarios individually, 
creating nine inundation maps. These show the physical floodplains of each scenario modeled, 
which are the extents used to determine vulnerability. However, these nine different maps make it 
difficult to summarize which structures are actually experiencing the most vulnerability. To 
complement the initial single variable inundation maps, summary multi-hazard maps were created 
to more easily distinguished vulnerability within each of the communities using layered 
symbology.  
There were several techniques that could have been used to visualize multivariate 
symbology, such as trivariate choropleth maps, however, trivariate choropleth symbology is 
usually reserved for attributes that add up to 100%, such as votes or grain size composition 
(Slocum et al., 2005). Layered symbology, however, allows for separate attributes to be 
visualized in a more independent manner. Also, the trivariate choropleth maps have a stretched 
color scheme, rather than the classified color scheme used in the layered symbols, which does not 
allow the map reader to designate separate hazard vulnerabilities as well. 
 
27 
 
4.6 Vulnerability Assessment 
Visual information graphics and a scoring system are used to assess the vulnerability of 
each type of infrastructure and place. Each hazard for each study area will have both a map of 
modeled inundation, and an associated bar graph with it as well to give the relative vulnerability 
of each structure type. The scoring system is based on normalizing the number of vulnerable 
features to the total gross count of that system component. The absolute vulnerability will be 
noted and discussed, but will not be included on the bar graphs. These bar graphs then will assist 
in comparing the differing amount of vulnerability experienced by each hazard for all three 
communities.  
4.7 Hypsometric Graphs 
Hyposometric graphs, which plot water level elevation versus the area inundated, are 
developed for each community and evolving time step of future risk. These graphs were 
developed in order to show how vulnerability may increase over time, as well as to reveal how 
elevation-area thresholds in the different hazards are possible where break points in flooded areas 
can occur at specific elevations of the water level (Zhang et al., 2011). If higher elevations occur 
closer to the water source, but lower elevations exist behind this peak, once water has breached 
this point, it will suddenly flood everything lower behind it. These graphs aid in the assessment of 
how the topography affects the threat, and how the elevation gradient in a community affects 
flooding (Figure 4.4).  
A steeper slope on these graphs represent a sudden increase in flood extents over a small 
increase in flood elevation, which corresponds to a gradual sloping, or flat topography. A gradual 
slope on the graph means there is very little change in the areal extent of the floodplain, while 
flood elevations continue to rise, which corresponds to a steeper topographic surface. Each study 
site has one graph showing this relationship for only storm surge and sea level rise, and does not 
28 
 
include the riverine flood hazard. The rationale for this is that flood height has a different 
relationship with flood area for riverine floods since it is water runoff, draining towards the coast, 
rather than water rising from the coast, such as storm surges or sea level rise.  
 
 
Figure 4.4: An example hypsometric graph with annotation to demonstrate how different slopes in 
the relationship of flood height and flood area help to illustrate coastal topography. 
 
 
 
 
 
 
CHAPTER 5: RESULTS 
 The infrastructure for all three study sites is shown in figures 5.1, 5.2, 5.3, and the 
population served by wastewater systems found through dasymetric mapping is shown in table 
5.1. As expected, New Bern, the only study site classified as a city rather than a town, has the 
highest population, the most treatment plants, most pump stations, the longest length of pipelines, 
and is the only locality with booster stations. Booster stations and pump stations are similar 
structures that are placed along pipelines that help move the fluid along the pipeline. Specifically, 
booster stations are for clean water pipelines, while pump stations are for wastewater. It should be 
noted that New Bern has four registered booster stations in the city, but one of them is currently 
offline, according to a local engineer who published the infrastructure data. It was thus excluded 
from this study. The population served, pump stations, and length of pipelines are an entire order 
of magnitude higher in New Bern than in Plymouth or Manteo. Manteo and Plymouth have much 
more comparable numbers, both containing only one treatment plant of each kind, but Plymouth 
possesses twice the number of sewage pump stations, and serves about 1,000 more people than 
Manteo.  Population serviced by wastewater utilities were sometimes higher than the population 
of the city or town, which could be due to either the low precision of the dasymetric population 
distribution estimates, or that people serviced by the treatment plant live outside of the city or 
town limits. 
 
 
 
 
 
 
30 
 
Table 5.1: Water infrastructure and population served by wastewater services for all three study 
sites. Population within city/town limits provided by 2010 Census (Census, 2011).  
 
31 
 
 
Figure 5.1: Water infrastructure for Manteo, NC overlaid onto orthophotography (composited of 
2010, 2011 and 2013 imagery). 
32 
 
 
Figure 5.2: Water infrastructure for New Bern, NC overlaid onto orthophotography (composited 
of 2010, 2011 and 2013 imagery). 
33 
 
  
Figure 5.3: Water infrastructure for Plymouth, NC overlaid onto orthophotography (composited 
of 2010, 2011 and 2013 imagery). 
34 
 
Inundation maps are paired with bar graphs that show where inundation takes place 
spatially, along with how the infrastructure is vulnerable on a normalized basis, as explained 
previously. Each study site, Manteo, New Bern, and Plymouth, has separate maps for each 
hazard. Riverine flooding is in figures 5.4, 5.5, and 5.6, storm surge in figures 5.7, 5.8, and 5.9, 
and sea level rise in figures 5.10, 5.11, and 5.12, respectively.  
5.1 Riverine Flooding 
Riverine flooding creates the greatest vulnerability for both Plymouth and Manteo, and the 
second most for New Bern. While Plymouth experiences the most vulnerability from riverine 
flooding than it has from either of the other two coastal hazards, none of the hazards in this study 
put the clean water treatment plant at risk. The 100 year floodplain puts all of the types of 
infrastructure at risk, except for the aforementioned clean water treatment plant, with the clean 
and wastewater pipelines, wastewater pump stations and wastewater treatment plant at 9.2%, 
18.5%, 35.7% (5 of 14), and 100% (1 of 1) vulnerability respectively. The 500 year floodplain 
increases the vulnerability of the clean and wastewater pipelines and pump stations to 18.7%, 
33.4%, and 57.1% (8 of 14) respectively. 
New Bern has the second greatest infrastructure vulnerability from riverine flooding, with  
all of the types of infrastructure at risk within the 100 year floodplain, except for the clean water 
treatment plant, which, like Plymouth, was not at risk by any of the modeled hazards in this study. 
The 100 flood put New Bern at about 25% (within 3%) vulnerability for all of these types of 
infrastructure, except for the booster stations, which are at 33.3% (1 of 3) vulnerability. The 500 
year floodplain only increased the vulnerability of the the clean and wastewater pipelines to 
30.4% and 27.6% respectively, and the wastewater pump stations to 30.8% (32 of 104). The 
35 
 
wastewater treatment plant vulnerability did not increase, staying at only one inundated plant, but 
one more booster station became inundated, increasing the vulnerability from 33.3% to 66.7% (2 
of 3). 
Manteo experiences the greatest vulnerability from riverine flooding, with 100% (1 of 1) 
vulnerability for both treatment plants, and all seven pump stations, as well as 95% for 
wastewater lines and 82.7% in clean water lines, just in the 100 year floodplain. The 500 year 
floodplain slightly increases the vulnerability for both pipelines, from 95% to 98.3% and 82.7% to 
84.8% for wastewater and clean water pipelines respectively. 
36 
 
 
Figure 5.4: Riverine flooding vulnerability for Manteo, NC using FEMA modeled 100 and 500 
year floodplains. 
37 
 
 
Figure 5.5: Riverine flooding vulnerability for New Bern, NC using FEMA modeled 100 and 500 
year floodplains. 
38 
 
 
Figure 5.6: Riverine flooding vulnerability for Plymouth, NC using FEMA modeled 100 and 500 
year floodplains. 
39 
 
5.2 Storm Surge 
Storm surge creates the second highest vulnerability for Manteo, reaching 100% (1 of 1) 
vulnerability for both treatment plants, and pump stations for a category 5 hurricane, which also 
accounts for 94.2% and 80.8% vulnerability in wastewater and clean water pipelines respectively. 
While the pump stations and wastewater treatment plant do not reach 100% vulnerability until a 
category 5 hurricane, the clean water treatment plant reaches 100% with a category 2. Also, with  
a category 1 hurricane, pump stations experience 42.9% (3 of 7) vulnerability, wastewater 
pipelines 17.5%, and clean water pipelines 15.3% vulnerability. 
New Bern experiences the greatest vulnerability from modeled storm surge than it does 
from either of the other two coastal hazards in this study. A category 1 storm put three types of 
infrastructure at risk, with clean and wastewater pipelines at 7.5% and 8% vulnerability 
respectively, and wastewater pump stations at 4.8% (5 of 104) vulnerability. A category 3 
hurricane will put all of the other types of water infrastructure at risk, excluding the clean water 
treatment plant mentioned above, inundating 25% (1 of 4) of the wastewater treatment plants, 
increasing the vulnerability of clean water and wastewater pipelines, and wastewater pump 
stations to 33%, 32.2% and 35.6% (37 of 104) respectively, and inundating 66.7% (2 of 3) of 
water booster stations. A category 5 hurricane will theoretically create the same vulnerability for 
wastewater treatment plants and water booster stations, but will increase the vulnerability of clean 
water and wastewater pipelines and wastewater pump stations to 54%, 49.1% and 50% (52 of 
104)  respectively. 
Modeled storm surge accounts for the second highest vulnerability for Plymouth, with a 
category 1 hurricane only inundating clean and wastewater pipelines and pump stations, 
generating a vulnerability for those structures of 0.1%, 3.8%, and 7.1% (1 of 14) respectively. The 
40 
 
only wastewater treatment plant becomes inundated with a category 3 hurricane, which also 
increases the vulnerability of clean water pipelines to 3.1%, wastewater pipelines to 10.9% and 
wastewater pump stations to 21.4% (3 of 14). A category 5 hurricane theoretically will only 
increase the vulnerability of the clean and wastewater pipelines to 3.8% and 12.1%, respectively. 
While all five categories were modeled in this study, it should be noted that a category 5 
hurricane is exceedingly unlikely for North Carolina given the total energy requirement to 
generate such a storm and the latitude of this area. However, flooding can be exacerbated through 
compound flooding, such as simultaneous rainfall and rivereine discharge, which may add to the 
surge generated in the SLOSH model. Antecedent flooding may also increase flood extents as 
well, such as a flood event affecting the area and not returning to normal water levels before a 
hurricane strike. For example, the combination of these two effects is the amplified Hurricane 
Floyd’s 500 year-flood event in 1999, despite being only a category 1 hurricane. While a category 
5 hurricane is an unlikely scenario for North Carolina to experience, the flood extents modeled by 
it in this study could be reached through compounded or antecedent flooding.  
 
41 
 
 
Figure 5.7: Modeled storm surge flooding for Manteo, NC generated from SLOSH MOM outputs 
downscaled to the local LiDAR DEM (6.1 m). 
42 
 
 
Figure 5.8: Modeled storm surge flooding for New Bern, NC generated from SLOSH MOM 
outputs downscaled to the local LiDAR DEM (6.1 m). 
43 
 
 
Figure 5.9: Modeled storm surge flooding for Plymouth, NC generated from SLOSH MOM 
outputs downscaled to the local LiDAR DEM (6.1 m). 
44 
 
5.3 Sea Level Rise 
Sea level rise created the least amount of vulnerability for all three study sites. Manteo 
reaches 100 % (1 of 1) vulnerability for its wastewater treatment plant at 1 meter, and its clean 
water treatment plants at 80 cm of sea level rise. A rise in 1.5 m of sea level results in 60.3% 
vulnerability for clean water pipelines, 65.3% vulnerability for wastewater pipelines, and 71.4% 
(5 of 7) for pump stations. 
Modeled sea level rise accounts for the least vulnerability in New Bern for water 
infrastructure, only inundating three types of structures: clean and wastewater pipelines and 
wastewater pump stations.  These three structures are at risk for all five sea level rise scenarios, 
and at 40 cm of sea level rise, clean and wastewater pipelines, and wastewater pump stations 
show 3.8%, 3.5%, and 1% (104) vulnerability respectively. A rise in 1.5 m of sea level increases 
their vulnerability to 14.4%, 14.3% and 8.7%  respectively. 
Finally, in Plymouth, sea level rise only inundated the wastewater treatment plant at 1.5 
m, which also puts the clean and wastewater pipelines at 1.6% and 7.6% vulnerability 
respectively, and wastewater pump stations at 14.3% (2 of 14) vulnerability. The wastewater 
pump stations first experience vulnerability at 80 cm of rise, with only 7.1% (1 of 14), which 
stays constant until doubling (2 of 14) at 1.5 m.  
 
45 
 
 
Figure 5.10: Modeled sea level rise flood potential for Manteo, NC. 
46 
 
 
Figure 5.11: Modeled sea level rise flood potential for New Bern, NC. 
47 
 
 
Figure 5.12: Modeled sea level rise flood potential for Plymouth, NC.
48 
 
 
To assist in comparing the vulnerability across both hazards and communities, a small 
mutliples technique was used to display a comprehensive diagram to show all nine scenarios 
(Figure 5.13). This figure was organized so that vulnerability decreased from the top left to 
bottom right, and helped to distinguish how vulnerability varied across the board. 
 
Figure 5. 13: Summary chart of all nine flood scenarios. This figure assists in comparing how 
vulnerability varied among the communities and hazards. 
49 
 
5.4 Hypsometric Graphs 
5.4.1 Manteo 
The hypsometric graph for Manteo (Figure 5.14) shows that storm surge reaches both a 
higher flood height, at 1.97 m, and creates a larger flood area, at 15.65 km2, than sea level rise, 
which reaches a flood height of 1.57 m and creates a flood area of 13.36 km2. The relationship for 
storm surge between flood height and area follows a pattern closer to a linear relationship, 
gradually increasing for both flood height, 0.7 m to 1.97 m, and flood area, 9.43 km2 to 15.65 
km2. Sea level rise, however, increases gradually only for the flood height, 0.68 m to 1.57 m, but 
the flood area has a sharp increase initially from 0.54 km2 to 9.72 km2, and then gradually 
increases from there to 13.36 km2.  
 
Figure 5.14: Modeled sea level rise and storm surge inundation and areas for Manteo, NC. 
50 
 
5.4.2 New Bern 
New Bern has much larger differences between the sea level rise and storm surge 
hypsometric curves (Figure 5.15). Storm surge far surpasses sea level rise in both flood height and 
area, at 5.36 m of flood height which covers 191.28 km2 for a category 5 storm surge, compared 
to sea level rise, which reaches a maximum 2 m flood height, covering 68.79 km2 for a 1.5 m sea 
level rise scenario. Both hazards have a similar linear shape of gradually increasing flood area 
with an increase in flood height, but storm surge increases in both of these categories much more 
quickly than sea level rise. A category 1 storm surge generated a 1.3 m flood height which 
covered 54.79 km2, while a 40 cm sea level rise scenario generated a 1.2 m flood height which 
covered 38.77 km2. 
 
Figure 5.15: Modeled sea level rise and storm surge flood heights and areas for New Bern, NC. 
  
51 
 
5.4.3 Plymouth  
The Plymouth hypsometric curves (Figure 5.16) have very similar shapes for both sea 
level rise and storm surge, with storm surge generating larger flood heights and flood area at 1.76 
m and 11.95 km2 for a category 5 hurricane, while a 1.5 m rise in sea level generates a flood 
height of 1.28 m which covers 10.74 km2. Both hazards started with very similar flood height and 
extents, with 40 cm of sea level rise generating a 0.59 m flood height and 0.70 km2 flood area, 
and a category 1 hurricane creating a 0.76 m flood height and 0.65 km2 flood area. Both hazards 
went on to change very little in flood area with the next scenario, but also both show a sharp 
sudden increase in flood area, and eventually tapering off to a gradual increase again. For the 
storm surge scenarios, the sudden change came between the category 2 and category 3 hurricane, 
with a category 2 storm flooding 0.71 km2, and a category 3 storm flooding 11.46 km2. For the 
sea level rise scenarios, the sudden change occurred between 80 cm and 1 m of sea level rise, with 
80 cm of rise covering on 0.92 km2, and 1 m of rise covering 9.68 km2.  
 
Figure 5.16: Modeled sea level rise and storm surge flood heights and areas for Plymouth, NC. 
52 
 
5.5 Multi-hazard Analysis 
 Summary maps were created to aid visual interpretation of vulnerabilities from all three 
hazards within a single map for all of the structures, excluding pipelines (Figure 5.18, 5.19, 5.20). 
Preceding these three figures is an example multi-hazard map with annotated text boxes to help 
interpreting the complex symbology (Figure 5.17). Pipelines were excluded since the layered 
symbology used would generate too much noise on the map, and degrade the legibility. Each 
symbol represents one structure, with up to three different vulnerability measurements, one for 
each hazard, with the center symbol representing riverine vulnerability, the middle symbol 
representing sea level rise vulnerability, and the outside symbol representing storm surge 
vulnerability. Color is used to visually represent the lowest scenario at which the structure could 
experience potential inundation. While all three hazards are represented, some of the structures 
are only vulnerable to one or two of the hazards rather than all three, so these points do not feature 
the layer that does not pose a risk for that structure.  
   
53 
 
 
Figure 5.17: An example of the following multi-hazard maps, with explanations on how to 
interpret the symbology. Each type of structure is explained, with instructions on how to interpret 
all three layers of the symbol. 
54 
 
 
Figure 5.18: Multi-hazard vulnerability from modeled sea level rise, storm surge, and riverine 
flooding for Manteo, NC. Manteo contains the only clean water treatment plant of all three study 
sites, and all seven of their pump stations experience vulnerability at some point. 
55 
 
 
Figure 5.19: Vulnerability from modeled sea level rise, storm surge, and riverine flooding for 
New Bern, NC. New Bern has the highest count of vulnerable infrastructure, and has the only 
booster stations of all three study sites, all of which experience some vulnerability. 
56 
 
 
Figure 5.20: Vulnerability from modeled sea level rise, storm surge, and riverine flooding for 
Plymouth, NC. Plymouth’s waterfront wastewater treatment plant is placed in a highly vulnerable 
area, as well as several pump stations placed along the Conaby Creek floodplain. 
 
 
CHAPTER 6: DISCUSSION AND CONCLUSIONS 
6.1 DISCUSSION 
 Manteo generated the highest vulnerability percentages for its infrastructure across all 
three coastal hazards, making it the most vulnerable community in this study. All of the treatment 
plants and pumping stations potentially become inundated at some point in both the storm surge 
and riverine flood simulations, and it also has the only clean water treatment plant at risk of all 
three communities. Losing these treatment plants would not only be very costly to this 
community, but would also cut off the ability to flush sewage and would make fresh water 
inaccessible for the people of Manteo. Manteo experiences vastly larger flood extents than either 
of the other communities in response to such low changes in the water level. To explain Manteo’s 
extensive flood vulnerability, the combination of a large area of low elevation, a gently sloping 
coastline, and being a coastal estuarine island, which allows flooding to occur on both the eastern 
and western shores of Roanoke Island, are important contributing factors.  
 Overall, New Bern is the second most vulnerable study site of the three, exceeding 
Plymouth in relative vulnerability in almost every category except for wastewater treatment 
plants, owing to one of four plants being at risk for New Bern, versus the sole plant serving 
Plymouth being at-risk. New Bern also certainly experiences the most absolute vulnerability of all 
three communities, with 53 pump stations at risk of at least one of the three hazards, and all three 
booster stations.  In addition, while New Bern’s clean water treatment plant does not apparently 
become at-risk, it is the only site that requires water booster stations to distribute fresh water to 
the community, and some of these structures experience inundation in some scenarios. This is an 
example of indirect, or cascading, vulnerability since losing the use of these booster stations will 
induce similar effects to losing the use the treatment plant itself. So, New Bern also has some 
58 
 
population that is vulnerable to the loss of both of the ability to flush sewage and inaccessibility to 
fresh water. New Bern is also distinct from the other communities in that it covers a much larger 
area, so the potential flood inundation extents are much larger. Storm surge prone areal extent, for 
example, is an order of magnitude larger than any of the other hypsometric curves, and flood 
heights are over 3 meters higher than any of the others as well. In addition to being a larger study 
area, as hypothesized, New Bern’s estuarine situation leads to these larger flood areas and heights. 
These curves illustrate that New Bern’s coast has a gradual slope, which exacerbates New Bern’s 
vulnerability.  
 Lastly, Plymouth’s water infrastructure is the least vulnerable of all of the sites. While the 
wastewater treatment plant experiences some imminent vulnerability, it is the only site that does 
not experience vulnerability to either clean water treatment plants or booster stations. This means 
that, during a flood event such as a hurricane or heavy rainfall, there is no modeled risk of losing 
access to fresh water for the community of Plymouth. There are, however, some areas of clean 
water pipelines that are vulnerable to flooding. However, they experience a more long-term threat 
of failure, rather than failing from a single event. The hypsometric curves show that Plymouth 
initially has a steeper coast, but is surrounded by a shelf of flat land, that tapers off to higher 
ground. As hypothesized, Plymouth experiences the least vulnerability and flood area from storm 
surge than any of the other sites, which is mostly a function of its sheltered riverine situation just 
inland from the Albemarle Sound and the narrow mouth of the Roanoke River.  
 The ranking of these communities on their vulnerability is a difficult task due to the 
complexity of having both an absolute and relative vulnerability. Manteo does seem to be the 
most overall vulnerable of all three communities based solely on its extremely high relative 
vulnerability. However, both Manteo and Plymouth have the same number of at risk structures 
59 
 
counting pump stations and treatment plants, with Manteo having slightly longer lengths of 
pipelines at risk. It is the combination of Manteo’s vastly high relative vulnerability coupled with 
having both types of treatment plants at risk as well that make it considered the most vulnerable 
community.  Similarly, it is the cumulation of these considerations that New Bern was ranked as 
the second most vulnerable compared to Plymouth, despite the 100% relative vulnerability that 
Plymouth experiences for its wastewater treatment plant. Plymouth’s both low relative and 
absolute vulnerability, along with  the lack of vulnerability to their clean water infrastructure 
makes them the least vulnerable.  
6.2 Limitations and Improving Coastal Inundation Studies 
 There are several changes that could be made to this study to improve comparative coastal 
vulnerability assessments of flooding. First, the method used to determine flood heights within the 
hypsometric graphs were not easily obtainable through ArcGIS, so we created a methodology 
here that seemed to vary in accuracy with our own inputs. Ideally, for the flood heights in the 
hypsography for sea level rise, they would be equal to the input of sea level rise in the study (i.e. 
40 cm, 60 cm, 80 cm, 1 m, and 1.5 m), however, the storm surge water levels were not always 
input as one value for the entire study site. Zonal statistics of the DEM were used as a proxy to 
estimate the highest elevation value within the DEM that was inundated. Then, to ensure 
comparability and consistency of methodologies to obtain these values, the same procedure was 
used with the sea level rise inundation models as well. In some cases, the output of this method 
for sea level rise was very similar to the inputs of sea level rise that we used, but in other 
instances, it may have overestimated inundation. So, the flood height values themselves may not 
be ideally reliable, however, they still graphically describe the elevation of the coast lines of these 
communities quite well. Verification stems from the fact that both sea level rise and storm surge 
60 
 
hypsographs for Plymouth show the same shape of a steep coast line, followed by a large, flat 
shelf, and finally a gently sloping topography further away from the shoreline on the Roanoke 
River. Additionally, if a smaller storm surge flood height, such as a tropical storm, were modeled 
for Manteo, a similar drastic decline in flood area would be seen as the flood height decreased, 
much like the one seen from 60 cm of sea level rise to 40 cm.  
 For resolution consistency, it would also be fruitful to match the resolution of the models 
used to generate the DFIRM outputs (50 ft. / 15.25 m) with the resolution of the sea level rise and 
storm surge models (20 ft. / 6.10 m). This could be done by either running a dedicated hydraulic 
model, possibly the Hydrologic Engineering Centers River Analysis System  (HECRAS) or 
another spatially explicit hydrologic model, or pursuing downscaling of existing models with 
similar methods to how the SLOSH storm surge model output was used. The wastewater 
treatment plant in Plymouth along the Roanoke River is on elevated ground which is seen by the 
higher resolution models, but it is uncertain if this treatment plant is truly in the 100 year 
floodplain, or if the small area is not represented well in the lower resolution DFIRM model. 
Other issues with misrepresentation may lie within the DEM itself. There will always be some 
error within the elevation values from the LiDAR point cloud, which is continually being 
improved as technology improves. Also, there may as well be adaptive structures already in place, 
such as berms or elevated platforms, that are not captured by the bare earth interpolated surface. 
Due to these issues, prior to any adaptive measures, there should be finer scale, individualized 
studies to verify the DEM for accuracy.  
 Perhaps the largest limitation in this study is the isolation of these hazards, rather than the 
ability to show compounding or antecedent flooding effects. All of the models in this study 
assume normal water levels, and base inundation by comparing a rise in water level with the 
61 
 
DEM. However, some flood events, such as that caused by Hurricane Floyd, have been shown to 
be far under-predicted because of heavy rainfall, and water tables that were still high from recent 
preceding storms, which are factors that are not included in our models. In addition, as sea level 
rises, water levels in these communities will be permanently higher, making 100 and 500 year 
floods much more frequent, and will likely make floodplains more extensive. Sea level rise also 
creates other future unknowns, such as how it will reshape the coast, and how that will affect the 
severity of coastal flooding. Rising water levels could eventually over-top the outer bank islands, 
which help protect the other coastal counties from large waves and tides. Without these islands, 
storms could generate larger surges due to larger fetches. With climate change and sea level rise 
models possessing so much uncertainty, mitigation efforts such as reduction in carbon emissions 
can help in ensuring a safer future for our coastal communities. 
6.3 Contributions of This Research 
It is hoped that this study will bring more attention to the dangers of coastal hazards, and 
to foster more progressive attitudes towards climate change and sea level rise. This approach of 
comparing hazards separately should inform coastal hazards researchers about the varying effects 
(spatially and temporally) of flood hazards in coastal communities. The vulnerability of this 
infrastructure is significant to the entire region, and more communities should be looking into the 
future to keep their citizens and environment in the healthiest condition as possible. Since there is 
a need for much of this infrastructure to be underground, its vulnerability will always be a 
complex issue, but the importance of thinking ahead for solutions where there is a clear and 
present danger is great. Many climate change projections reveal serious challenges for the United 
States’ coasts. Assessing and preparing for these increasing hazards will hopefully lead to fewer 
62 
 
distasters, less costly response and recovery efforts, and a safer environment for coastal 
communities in the long run. 
 Much of this study has a clear basis using GIS for modeling and interpretations. While 
GIS is certainly a powerful tool that has some striking abilities to convey complex messages, it 
should be used with caution and a fine attention to detail. Some geographers are critical of how 
GIS has transformed the field of geography. For example, Peter Taylor believes that it drives a 
regression of knowledge from ideas to simply facts, calling it the very worst sort of positivism, 
and allowing geography researchers to ignore any broader questions in a social or political 
context (Taylor, 1990). However, studies like this one would be far more difficult without the 
powerful abilities of GIS, and in fact, geospatial information is critical to hazards identification, 
response and preparedness. With a strong background in attention to what geography is, an 
abundance of applied research can be empowered through its broad range of capabilities. 
Openshaw (1991) saw GIS for what it was early on, a way to bring geography to more people, 
and viewed it as a holistic approach to the space-time foundation of geography.  
 The rapid and accurate data processing of which GIS is capable of has much to contribute 
to the study of natural hazards. Hazards researchers like White, Kates, and Burton (2001) question 
why vulnerability to coastal floods has escalated, when our knowledge of how these hazards 
function has only increased. With our assistance in developing an understanding in coastal flood 
hazards, as well as encouraging mitigation efforts to decrease vulnerability, this project reaches 
the very core of geographic study. It brings together the physical behavior of coastal flooding, 
while focusing on the human impacts of hazards. This combination provides the necessary tools 
for future natural hazard management and adaptation. 
63 
 
6.4 Conclusions 
 The multi-hazard maps created for this study were very important in demonstrating the 
necessity for the mutli-hazard approach. Some of the structures had only one or two layers 
present, meaning they were only vulnerable to certain hazards rather than all three.This indicates 
that other studies that only assessed vulnerability to a single hazard, such as storm surge, may not 
recognize that structure as having flood vulnerability. By analyzing all three hazards, we have 
obtained a more full assessment of flood vulnerability to this infrastructure, and thus are 
approaching the reduction of this hazard in a more complete manner. By the color scheme and 
layering technique used, the map reader can get a quick assessment of the overall vulnerability to 
each structure by the number of layers and color hue. Additionally, using three different study 
areas allowed for supplementary context for how flood hazards differ in changing environmental 
scenarios.  
 While reviewing their vulnerability to coastal floods, communities studied here will have 
to decide on the steps they want to take to mitigate and adapt based on the benefits it can bring 
them and the cost it will take to make changes. Manteo may have a harder time adapting to 
change than the other two sites due to a lack of land space that would create a less vulnerable 
scenario. Since the topography of Manteo is lower and flatter than either Plymouth or New Bern, 
moving structures, such as a treatment plant, is a very limited option, and, since moving an entire 
treatment plant would be costly procedure, the new placement of that plant would have to 
outweigh those moving costs with enough benefits to justify it. Hence, mitigative engineering 
approaches, such as elevating and protecting the structures, may be the only short to medium 
range options. In the longer term, even beyond the scope of the scenarios included in this study, 
64 
 
relocation and retreat alternatives are conceivable for Manteo and possibly fundamental to the 
community’s sustainability.   
 New Bern and Plymouth, however, have more room to explore adaptation options. Both of 
these towns should weigh the options in moving their wastewater treatment plants that are at risk, 
because doing so would drastically decrease their overall vulnerability. Specifically, if Plymouth 
could successfully move their riverside wastewater treatment plant to adjacent higher elevated 
land, adaptation for several pump stations would almost eradicat any remaining coastal flood 
vulnerability except for pipelines. Pipeline adaptation strategies are different from pump station or 
treatment plant strategies, in that they can not simply be moved or raised, because their location is 
necessary to provide people with their service. However, there are other adaptation strategies, 
such as pipe coverings that protect from salt water corrosion, which may also help these 
communities reduce their vulnerability. Since New Bern has a much greater absolute vulnerability 
than the other communities, an emphasis on their adaptation strategies should be prioritizing more 
vulnerable structures in order to systematically lower their vulnerability over time. It is also 
notable that adaptation can accommodate risk reduction for multiple hazards, so the multi-hazard 
assessment here could assist the community with addressing long-term capital and adaptation 
planning.  
 While adaptation strategies should be weighed more on a cost-benefit analysis, mitigation 
through reduction of carbon emissions should be a goal for the entire country. Not only does it 
aim to make our communities more sustainable, but as sea level rise and climate change continue, 
their impacts are likely to worsen, and may be hard to predict. Sea level rise accounted for the 
least vulnerability in this study, but only because it was calculated as an isolated hazard, rather 
than being viewed as an added threat to storm surge or riverine flooding. Ultimately, accelerating 
65 
 
sea levels will drive all the other coastal hazards and changes, compounding storm surges, rising 
water tables, changing coastal geomorphology and ushering the migration of wetlands. Sea level 
rise is more of a threat than is captured in the modeling used here, and will become the test of 
sustainable communities for generation to come.  
 
 
 
References 
Allen, Tom R., Stephen Sanchagrin, and George McLeod. (2013). "Visualization for Hurricane 
Storm Surge Risk Awareness and Emergency Communication." Approaches to Disaster 
Management: Examining the Implications of Hazards, Emergencies and Disasters. Rijeka: 
Intech, 2013. 105-30. Print. 
Aycock, W.C., Wang, Y. (2004). Comparison of the New Digital Flood Insurance Map (DFIRM) 
With the Existing FIRM, Wilson, NC. Southeastern Geographer, 44(2), 159-169.  
Bales, J.D. (2003). Effects of Hurricane Floyd Inland Flooding, September – October 1999, on 
Tributaries to Pamlico Sound, North Carolina. Coastal and Estuarine Research 
Foundation, 26 (5), 1319-1328. 
Beven II, J., Avila, L., Blake, E., Brown, D., Franklin, J., Knabb, R., Pasch, R., Rhome, J., 
Stewart, S. (2007). Atlantic Hurricane Season of 2005. Monthly Weather Review, 136, 
1109-1173. 
Buhr, W. (2003). What is infrastructure? Discussion Paper No. 107-03, Department of 
Economics, School of Economic Disciplines, University of Siegen, Siegen.  
Burkett, V.R. and Davidson, M.A. (2012). Coastal Impacts, Adaptation and Vulnerability: A 
Technical Input to the Climate Assessment. Cooperative Report to the 2013 National 
Climate Assessment. 150.  
Cazenave, A., Llovel, W. (2010). Contemporary Sea Level Rise. Annual Review of Marine 
Science. 2, 145-173.  
Chughtai, F., Zayed, T. (2008). Infrastructure Condition Prediction Models for Sustainable Sewer 
Pipelines. Journal of Performance of Constructed Facilities, 22, 333-341. 
 
 
Clough, J.S., Park, R.A., and Fuller, R. (2009). The Nature Conservancy: SLAMM 6 Technical 
Documentation. Retrieved from 
http://warrenpinnacle.com/prof/SLAMM6/SLAMM6_Technical_Documentation.pdf  
Davidson-Arnott, R. (2005). Conceptual Model of the Effects of Sea Level Rise on Sandy Coasts. 
Journal of Coastal Research, 21(6), 1166-1172. 
Davis, R., Dolan, R. (1993). Nor’easters.  American Science, 81, 428–39. 
Duke, G., Kienzle, S., Johnson, D., Byrne, J. (2003). Improving Overland Flow Routing by 
Ancillary Road Data Into Digital Elevation Models. Journal of Spatial Hydrology. 3, 1-27. 
Friedrich, E., Kretzinger, D. (2012). Vulnerability of Wastewater Infrastructure of Coastal Cities 
to Sea Level Rise: A South African Case Study. Water SA, 38(5), 755-764. 
Fugro. (2012). City of Norfolk City-wide Coastal Flooding Study. Retrieved from 
http://ccrm.vims.edu/recurrent_flooding/Recurrent_Flooding_Study_web.pdf  
Fink, F. W. (1960). Corrosion of metals in sea water. US Department of Commerce, Business and 
Defense Services Administration. 27-39. 
Heberger, M., Cooley, H., Herrera, P., Gleick, P., Moore, E. (2011). Potential Impacts of 
Increased Coastal Flooding in California Due to Sea-Level Rise. Climatic Change, 109, 
229-249. 
Hecht, J. (2006). Disappearing Deltas Could Spell Disaster. New Scientist, 189, 8–11. 
Hondula, D., Dolan, R. (2010). Predicting Severe Winter Coastal Storm Damage. Environmental 
Research Letters, 5(3), 34004. 
IPCC (International Panel on Climate Change) (2007) Chapter 6 – Coastal Systems and Low-
Lying Areas. In: Parry ML, Canziani OF, Palutikof JP, Van der Linden PJ and Hanson CE 
(eds.) Contribution of Working Group II to the Fourth Assessment Report of the 
 
 
Intergovernmental Panel on Climate Change, 2007. Cambridge University Press, Cambridge, 
New York. Retrieved from  
http://www.ipcc.ch/publications_and_data/ar4/wg2/en/contents.html  
Jelesnianski, C.P., Chen, J., Shaffer, W.A. (1992). SLOSH: Sea, Lake and Overland Surges from 
Hurricanes. Washington, DC: National Oceanic and Atmospheric Administration.  
Kearns, J., Pak, S., Buhayar, N. (2012). Sandy Seen Boosting U.S. with as Much as $240 Billion 
Rebuilding. Bloomberg News. Retrieved from http://www.bloomberg.com/news/2012-11-
23/sandy-seen-boosting-u-s-with-as-much-as-240-billion-rebuilding.html  
Kelly, M., Munson, A., Morales, J., Leeper, D. (2007). Proposed Minimum Flows and Levels for 
the Upper Segment of the Braden River, from Linger Lodge to Lorraine Road. The 
Southwest Florida Water Management District. Retrieved from 
http://www.swfwmd.state.fl.us/projects/mfl/reports/braden_mfl_peer_review.pdf   
Kemp, A., Horton, B., Culver, S., Corbett, D.R., van de Plassche, O., Gehrels, W.R., Douglas, B., 
Parnell, A. (2009a). Timing and magnitude of recent accelerated sea-level rise (North 
Carolina, United States). Geology, 37(11), 1035-1038. 
Kemp, A., Horton, B., Corbett, D. R. , Culver, S., Edwards, R., van de Plassche, O. (2009b). The 
Relative Utility of Foraminifera and Diatoms for Reconstructing Late Holocene Sea- 
Level Change in North Carolina, USA. Quaternary Research, 71(1), 9-21. 
Knutson, T., McBride, J., Chan, J., Emanuel, K., Holland, G., Landsea, C., Held, I., Kossin, J., 
Srivastava, A., Sugi, M. (2010). Tropical Cyclones and Climate Change. Nature 
Geoscience, 3, 157-163. 
Kundzewicz, Z.W., Matczak, P. (2012). Natural Risks: Mitigation and Adaptation. Ecohydrology 
and Hydrobiology, 12(1), 3-8.  
 
 
Lane, K., Charles-Guzman, K., Wheeler, K., Abid, Z., Graber, N., Matte, T. (2013). Health 
Effects of Coastal Storms and Flooding in Urban Areas: A Review and Vulnerability 
Assessment. Journal of Environmental and Public Health, 1-13.  
Mallin, Michael A., Cahoon, Lawrence B., Toothman, Byron R., Parsons, Douglas C., McIver, 
Matthew R., Ortwine, Michelle L., Harrington, Renee N. (2007). Impacts of a Raw 
Sewage Spill on Water and Sediment Quality in an Urbanized Estuary. Marine Pollution 
Bulletin, 54, 81-88. 
Munaretto, S., Vellinga, P., Tobi, H. (2012). Flood Protection in Venice Under Conditions of Sea 
Level Rise: An Analysis of Institutional and Technical Measures. Coastal Management, 
40(4), 355-380.  
NC Coastal Resources Commission. (2010). NC Sea Level Rise Assessment Report. NC Science 
Panel on Coastal Hazards.  Retrieved from http://dcm2.enr.state.nc.us/slr/NC%20Sea-
Level%20Rise%20Assessment%20Report%202010%20-
%20CRC%20Science%20Panel.pdf .  
Nicholls, R., Hoozemans, F., Marchand, M. (1999). Increasing Flood Risk and Wetland Losses 
Due to Global Sea-Level Rise: Regional and Global Analyses. Global Environmental 
Change, 9, 69-87.  
NOAA. (2011). Irene By The Numbers. National Hurricane Center. Retrieved from 
http://www.noaa.gov/images/Hurricane%20Irene%20by%20the%20Numbers%20-
%20Factoids_V4_083111.pdf  
Norris, F. H., Stevens, S. P., Pfefferbaum, B., Wyche, K., Pfefferbaum, R. L. (2008). Community 
Resilience as a Metaphor, Theory, Set of Capacities, and Strategy for Disaster Readiness. 
American Journal of Community Psychology, 41, 127-150. 
 
 
NY.Gov 2100 Commission. (2013). Superstorm Sandy Recover. Retrieved from 
http://www.governor.ny.gov/2013/superstorm-sandy-recovery-rebuild.  
Openshaw, S. (1991). A View of the GIS Crisis in Geography, or, Using GIS to Put Humpty 
Dumpty Back Together Again. Human Geography: An Essential Anthology, J. Agnew, D. 
Livingstone and A. Rogers, eds., 675-685. Malden, MA: Blackwell. 
Overton, M., Grenier R., Judge, E., Fisher, J. (1999). Identification and Analysis of Coastal 
Erosion Hazard Areas: Dare and Brunswick Counties, North Carolina. Journal of  Coastal 
Resources, 28, 69–84. 
Park, R. A., Armentano, T.V., and Cloonan, C.L. (1986). Predicting the Effects of Sea Level Rise 
on Coastal Wetlands. In: J.G. Titus, eds. Effects of Changes in Stratospheric Ozone and 
Global Climate, vol. 4: sea level rise. Washington, DC: U.S. Environmental Protection 
Agency, 129–152. 
Parry, M., Canziani, O., Palutikof, J., Van Der Linden, P., Hanson, C. (2007). Climate Change 
2007: Impacts, Adaptation and Vulnerability. Contribution of Working Group II to the 
Fourth Assessment Report of the Intergovernmental Panel on Climate Change, Cambridge 
University Press, Cambridge.  
Poulter, B. and Halpin, P. (2008). Raster Modelling of Coastal Flooding from Sea-level Rise. 
International Journal of Geographical Information Science. 22(2), 167-182. 
National Weather Service. (2013). Event Summary- Hurricane Floyd, September 1999. Retrieved 
from http://www4.ncsu.edu/~nwsfo/storage/cases/19990915/  
Rice, K.C., Hong, B., Shen, J. (2012). Assessment of Salinity Intrusion in the James and 
Chickahominy Rivers as a Result of Simulated Sea-Level Rise in Chesapeake Bay, East 
Coast, USA. Journal of Environmental Management, 111, 61-69. 
 
 
Slocum, T.A., McMaster, R.B., Kessler, F.C., Howard, H.H. (2005). Bivariate and Multivariate 
Mapping. Thematic Cartography and Geographic Visualization, 2nd Ed. (341-359). Upper 
Saddle River, NJ: Pearson Education, Inc. 
Small, C., Nicholls, R. (2003). A Global Analysis of Human Settlement in Coastal Zones. Journal 
of Coastal Research. 19(3), 584-599.  
Smith, J., Dolan, R., Lins, J. (2006). Hurricane History of the North Carolina Outer Banks (USA), 
1586 to 2004. Shore and Beach, 74(3), 19-23. 
Taylor, P.J. (1990). GKS. Political Geography Quarterly, 9, 211-12. 
Titus, J., Kuo, C., Gibbs, M., LaRoche, T., Webb, M. K., Waddell, J. (1987). Greenhouse Effect, 
Sea Level Rise, and Coastal Drainage Systems. Journal of Water Resources Planning and 
Management, 113(2), 216-227. 
U.S. Census Bureau. (2011). American FactFinder Fact Sheet: Retrieved October 21st, 2013, 
from http://factfinder2.census.gov/faces/nav/jsf/pages/index.xhtml 
USGS. (2014). Hurricane Irene Storm Tide Mapper. Retrieved from 
http://wim.usgs.gov/stormtidemapper/stormtidemapper.html 
Vermeer, M., Rahmstorf, S. (2009). Global Sea Level Linked to Global Temperature. PNAS, 1-6. 
White, G.F.  Kates, R.W., Burton, I. (2001). Knowing Better and Losing Even More: The Use of 
Knowledge in Hazards Management. Environmental Hazards, 3, 81-92. 
Wolf, S., Hinkel, J., Hallier, M., Bisaro, A., Lincke, D., Ionescu, C., Klein, R. (2013). Clarifying 
Vulnerability Definitions and Assessments Using Formalization. International Journal of 
Climate Change Strategies and Management, 5(1), 54-70 
Zachry, B. C., Doggett, A. L., Kennedy, A. B., & Keller, T. J. (2012). Slosh Model Hindcast of 
Hurricane Irene (2011) Surge. AMS Confex, 40. 
 
 
Zhang, K., Dittmar, J., Ross, M., Bergh, C. (2011). Assessment of Sea Level Rise Impacts on 
Human Population and Real Property in the Florida Keys. Climatic Change, 107, 129 – 
146.  
Zhang, K., Xiao, C., Shen, J. (2008). Comparison of the CEST and SLOSH Models for Storm 
Surge Flooding. Journal of Coastal Research, 24(2), 489-499.