assessment of inundation risk from sea level rise and ... · assessment of inundation risk from sea...
TRANSCRIPT
Assessment of Inundation Risk from Sea Level Rise andStorm Surge in Northeastern Coastal National Parks
Angelica Murdukhayeva†, Peter August†, Michael Bradley†, Charles LaBash†, and NigelShaw‡
†University of Rhode IslandDepartment of Natural Resources Science1 Greenhouse RoadKingston, RI 02881, [email protected]
‡National Park ServiceNortheast Region15 State StreetBoston, MA 02109, U.S.A.
ABSTRACT
Murdukhayeva, A.; August, P.; Bradley, M.; LaBash, C., and Shaw, N., 0000. Assessment of inundation risk from sealevel rise and storm surge in northeastern coastal national parks. Journal of Coastal Research, 00(0), 000–000. CoconutCreek (Florida), ISSN 0749-0208.
Sea level rise and an increase in storm frequency and intensity are two major impacts expected to result from climatechange in coastal ecosystems. Coastal national parks have many low-lying areas that are at risk from inundationresulting from these impacts. To help park managers meet their goal of preserving resources, we developed amethodology to evaluate risk of inundation from sea level rise and storm surge at sentinel sites, areas of importance fornatural, cultural, and infrastructural resources. We selected the most recent, readily available, and appropriategeospatial tools, models, and data sets to conduct case studies of our coastal inundation risk assessments in twonortheastern coastal national parks—Cape Cod National Seashore, MA, and Assateague Island National Seashore, MD/VA. We collected elevation data at sentinel sites using real-time kinematic global positioning system (RTK GPS)technology. We used three modeling approaches: modified bathtub modeling; the Sea Level Affecting Marshes Model(SLAMM); and the Sea, Lake, and Overland Surges from Hurricanes (SLOSH) model to assess the likelihood ofinundation at sentinel sites. Cape Cod’s sentinel sites, which in many cases occurred in high-elevation settings, werefound to be less vulnerable to inundation than were Assateague Island’s sentinel sites, which were distributed in low-lying areas along the barrier beach island. This inundation risk assessment methodology can be applied to other coastalareas and to the same coastal parks at different times as more accurate elevation data sets and updated sea level riseprojections become available.
ADDITIONAL INDEX WORDS: Coastal elevations, LIDAR, GPS, modeling, climate change.
INTRODUCTIONIn coastal ecosystems, accelerated sea level rise and an
increase in storm frequency and intensity are two major
impacts expected to result from climate change (Ashton,
Donnelly, and Evans, 2008; Bender et al., 2010; Harvey and
Nicholls, 2008). In the next century, the rate of global sea level
rise is anticipated to be several times higher than that
measured over the past century (Cazenave and Nerem, 2004;
Church and White, 2006; Overpeck and Weiss, 2009; Pfeffer,
Harper, and O’Neel, 2008; Rahmstorf, 2007). The U.S.
northeast coast experiences a rate of relative sea level rise
greater than the global average because of substantial regional
variations in glacial isostatic adjustment effects and oceano-
graphic processes (Sallenger, Doran, and Howd, 2012; Tami-
siea and Mitrovica, 2011). Along the U.S. Atlantic coast, the
highest rates of subsidence occur from southern Massachusetts
to Virginia (Engelhart, Peltier, and Horton, 2011), and
predicted changes in ocean circulation, driven by climate
change, could potentially add meters of dynamic sea level rise
along the northeast coast during the next century (Hu et al.,
2009; Yin, Schlesinger, and Stouffer, 2009). Furthermore, the
frequency and extent of severe coastal storms are expected to
increase (Bender et al., 2010), and large surge levels may cause
significant damage to coastal infrastructure and alteration of
ecosystems (Irish et al., 2010; Kirshen et al., 2008; Lin et al.,
2010; McInnes et al., 2003).
Our objective was to explore how widely accessible data and
tools can be used for highly localized assessments of inundation
risk. As case studies, we modeled inundation risk from
predicted sea level rise and storm surge at 97 sentinel sites in
two coastal, northeastern U.S. national parks—Cape Cod
National Seashore (CACO), and Assateague Island National
Seashore (ASIS)—using a variety of data sources, predicted
scenarios, and physical models.
The definitions of two terms used throughout this article are
essential to understanding the project. Sentinel sites are
DOI: 10.2112/JCOASTRES-D-12-00196.1 received 2 October 2012;accepted in revision 25 January 2013; corrected proofs received13 March 2013.Published Pre-print online 2 April 2013.� Coastal Education & Research Foundation 2013
Journal of Coastal Research 00 0 000–000 Coconut Creek, Florida Month 0000
locations of natural or cultural resources or critical infrastruc-
ture of special importance to the National Park Service (NPS).
Examples include piping plover habitat, historic buildings, and
bridges. At risk refers to the likelihood of inundation because of
sea level rise or storm surge. However, some sites and
associated resources predicted to be at risk of inundation
would not necessarily be negatively affected by it. For example,
a building may be unaffected by a brief period of inundation
from storm surge. In other cases, natural or cultural resources
or infrastructure may be severely affected if inundated.
Therefore, risk (of inundation) does not equate to ‘‘impact’’ in
our study. In either case, however, an understanding of the
vulnerability of each sentinel site to inundation will allow park
resource managers to develop appropriate management and
mitigation strategies for these sites and areas of concern.
Coastal zone scientists have used a variety of quantitative
models to assess the effect of sea level rise and storm surge on
coastal inundation and flooding. Several hydrodynamic models
are used to model inundation from rising sea levels. Examples
include MIKE 21 and MIKE FLOOD (Sto. Domingo et al.,
2010), ADCIRC (Lin et al., 2010), LISFLOOD-FP (Lewis et al.,
2011), and ANUGA (Van Drie, Milevski, and Simon, 2010). The
U.S. Federal Emergency Management Agency (FEMA) relies
on a Geographic Information Systems (GIS) model, HAZUS-
MH, to estimate the physical, economic, and social impacts of
large-scale flooding events (Scawthorn et al., 2006a,b). Other
GIS-based methods have been applied as well (Brown, 2006;
Hennecke and Cowell, 2000). A thorough review of sea level
and storm surge inundation models is provided in Murdu-
khayeva (2012).
Parkwide and regional-scale assessments of sea level rise
and storm surge risk have been conducted previously in these
two national parks (CACO, ASIS). The Coastal Vulnerability
Index (CVI) technique was applied at Cape Cod National
Seashore (Hammar-Klose et al., 2003) and at Assateague
Island National Seashore (Pendleton, Williams, and Thieler,
2004). The CVI method combines a number of physical
variables to classify the relative vulnerability of 1.5-km
shoreline segments to sea level rise effects. Gutierrez,
Williams, and Thieler (2007) studied potential shoreline
changes from sea level rise along the United States. Mid-
Atlantic coast. The CVI and shoreline-change assessments
provide a useful and robust overview of coastal vulnerability at
landscape scales (1.5 km stretches of coast). Our study assesses
risk at specific locations, thus provides a more-resolute
evaluation of inundation.
All vulnerability models and methods rely on elevation data,
which are often limited in their vertical accuracy and cause large
ranges of uncertainty in the results (Gesch, 2009). The National
Elevation Dataset (NED) is the largest-scale, readily available,
topographic data set for the entire United States, but with a
vertical accuracy (root mean square error [RMSE]) of 2.44 m
(Gesch, 2007), it is of little value in assessing inundation risk at
specific sites. Our study relied on two sources of high-accuracy
elevation data for sentinel sites: Light Detection and Ranging
(LIDAR) data and real-time kinematic global positioning system
(RTK GPS) data. Our assessment also integrated output for
three sea level rise scenarios and four storm surge scenarios
from several modeling approaches. For sea level rise, we used a
‘‘bathtub inundation’’ model with tidal–orthometric datum
conversion and the Sea Level Affecting Marshes Model
(SLAMM). For storm surge, we used the Sea, Lake, and
Overland Surges from Hurricanes (SLOSH) model. The proba-
bilities of inundation at sentinel sites were calculated for each
modeled scenario, and an index was developed to estimate
overall inundation likelihood at each sentinel site. This overall
index of inundation likelihood was an additional tool for
prioritizing long-term management and climate change adap-
tation plans in the national parks.
METHODSA goal of this research was to develop a process to assess sea
level rise and storm surge inundation risk at specific locations
in coastal environments. We endeavored to use off-the-shelf
data and technology and readily available models, so the
process could be used by other coastal resource managers in
other locations in the United States. A synopsis of the data and
tools we chose and the sources of additional information for
them are presented in Table 1.
Table 1. Data, tools, and models used in this study and references for further information.
Data/Tool Source Citations
Elevation Experimental Advanced Airborne
Research Lidar (EAARL);
the 3di Technologies, Inc.
Digital Airborne Topographic
Imaging System II (DATIS II),
RTK GPS Surveys
Bonisteel et al., 2009; Bonisteel-Cormier et al.,
2010; Brock et al., 2007; MassGIS, 2005,
performed by Neil Winn (NPS ASIS),
Mark Adams (NPS CACO),
Michael Bradley (URI), and
Angelica Murdukhayeva (URI)
Tidal and orthometric datum values NOAA CO-OPS, NOAA VDatum 3.0 beta NOAA, 2007; NOAA, 2011
SLOSH (Sea, Lake, and Overland Surges
from Hurricanes) model
NOAA NWS Display Version 1.64a
(release date: June 2011)
FEMA, 2003; Jarvinen and Lawrence, 1985;
Jelesnianski, Chen, and Shaffer, 1992
Storm surge interpolation tool Applied Science Associates, Inc. Isaji and Knee, 2009
SLAMM
(Sea Level Affecting Marshes Model)
Warren Pinnacle Consulting, Inc. Version 6.0.1 beta Warren Pinnacle Consulting, 2010
Wetlands U.S. Fish and Wildlife Service National Wetlands Inventory USFWS, 2010
Upland land cover NOAA Coastal Change Analysis Program NOAA, 2006
Local accretion rates J. Lynch, personal communication; NPS, 2009
Journal of Coastal Research, Vol. 00, No. 0, 0000
0 Murdukhayeva et al.
Study AreasWe worked in two coastal parks (Figure 1): Cape Cod
National Seashore (CACO) and the Assateague Island National
Seashore (ASIS). The CACO is located on the outer (eastern)
portion of Cape Cod, Massachusetts. It encompasses 123 km of
ocean, marsh, and inner bay shorelines and contains a variety
of marine, estuarine, and upland, terrestrial ecosystems with
relatively varied topography for the coastal zone. The ASIS is
located on the Delmarva Peninsula and lies within the
boundaries of two states, Maryland and Virginia. It has 60
Figure 1. Regional map with locations of study areas: (A) Cape Cod National Seashore (CACO), and (B) Assateague Island National Seashore (ASIS).
Journal of Coastal Research, Vol. 00, No. 0, 0000
Sea Level Rise and Storm Surge in Northeastern National Parks 0
km of shoreline on an undeveloped barrier island with long
stretches of dunes, beaches, and marshes. These study areas
represented two different types of coastal ecosystem land-
scapes: a peninsular glacial moraine (CACO) and a barrier
island (ASIS).
Data SourcesSentinel sites are locations of natural, cultural, and infra-
structural resources of special importance to the NPS and were
identified by park managers. At CACO, the 63 sentinel site
locations included groundwater-monitoring wells, lighthouses,
visitor centers, and historical monuments and sites, e.g. the
Marconi Site and historic Marindin shoreline survey markers
(Marindin, 1891). At ASIS, sentinel sites were locations of 34
newly installed or frequently used geodetic survey markers,
which are critical for ongoing research on the island geomor-
phology. We measured the elevation of sentinel sites using RTK
GPS (Trimble R8 GNSS), which provided elevation estimates
with average vertical RMSE values of 3.5 cm for CACO and 0.7
cm for ASIS.
Digital elevation models (DEMs) were derived from LIDAR
data obtained by the U.S. Geological Survey (USGS) and the
NPS from aircraft-mounted laser sensors. LIDAR data are
typically accurate to 0.15–1 m (Gao, 2007). The horizontal
resolutions of CACO’s and ASIS’s DEMs were 1 m and 2.5 m,
respectively. The reported vertical accuracy of the DEMs for
CACO and ASIS was 15 cm RMSE. To validate these accuracy
estimates, we calculated the vertical RMSE using high-
accuracy (,2 cm vertical and horizontal error) ground-control
points that were available in each park. For CACO, the vertical
RMSE of the DEM was 0.53 m using 35 existing geodetic
control points. For ASIS, the vertical RMSE was 0.33 m using
1179 control points.
The SLAMM model requires a DEM layer, a slope layer
(derived from the DEM), a detailed land-use layer, parameters
for tidal ranges (National Oceanic and Atmospheric Adminis-
tration [NOAA], 2007), and if known, accretion rates for
marshes. The land-use layer was created by merging the
Coastal Change Analysis Program (C-CAP) data (NOAA, 2006)
and the National Wetlands Inventory (NWI) data (USFWS,
2010) and recoding them to SLAMM categories, as specified by
the SLAMM 6.0.1 Beta Users Manual (Warren Pinnacle
Consulting, 2010). The source data for NWI were aerial
photography (CACO in 1993 and ASIS in 1988), and the source
data for C-CAP were 2006 Landsat imagery. Some model
parameters, such as marsh accretion rates, were obtained from
Surface Elevation Table (SET) data (J. Lynch, personal
communication; NPS, 2009). When these and other parameters
(i.e. beach sedimentation rate, frequency overwash) were
unknown or unavailable, SLAMM’s default settings were used.
Using default parameter settings could strongly sway the rates
of inundation predicted by SLAMM (Scarborough, 2009), but
we were limited by data availability for our case study sites.
Sea Level Rise Scenarios and ModelsThree scenarios were selected to represent the current range
of sea level rise predictions for the year 2100: rises of 0.6 m
(IPCC, 2007), 1 m (Vermeer and Rahmstorf, 2009), and 2 m
(Pfeffer, Harper, and O’Neel, 2008). The scenarios represent
reasonable estimates of low, medium, and high predictions of sea
level rise for this region. For example, the Rhode Island Coastal
Resources Management Council has adopted sea level rise
scenarios in the range of 1–2 m to guide their planning policies
(Gregg, 2010). Elevation data set accuracies also affected our
choice of scenarios; based on the recommendations of NOAA
(2010), vertical accuracies were sufficiently high for ASIS that
inundation mapping could be performed for the scenarios: 0.6 m,
1 m, and 2 m. Only the 1 m and 2 m scenarios could be modeled
for CACO because of the amount of uncertainty in the LIDAR
DEM for this area (vertical RMSE¼0.53 m). These two scenarios
for CACO and all three scenarios for ASIS were modeled using
two different sea level rise models.
The first modeling approach, the bathtub model, involved
creating a planar water surface representing the sea level rise
scenario that is added to the Mean Higher High Water
(MHHW) tidal elevation. Modeling sea level rise or storm
surge beyond the MHHW tidal elevation represents the worst-
case inundation scenario. These water surface elevations were
calculated using VDatum software (NOAA, 2011), which
performs elevation conversions between NAVD88 (an ortho-
metric datum) and tidal datums.
The second approach used to evaluate sea level rise
inundation was the Sea Level Affecting Marshes Model
(SLAMM). Using linear relationships and decision-tree rules,
SLAMM calculates water elevation and computes inundation
and habitat response over large areas. The SLAMM incorpo-
rates the temporal rate of the sea level changes (Mcleod et al.,
2010). The output maps show expected habitat classes and
areas of inundation based on the different rates and magni-
tudes of sea level rise associated with each modeled scenario.
Sentinel-site locations were mapped over the output, and a
change matrix was created for each one to show the shift from
initial land-cover class to the final, predicted land-cover class.
Storm Surge Scenarios and ModelsThe four selected scenarios for storm surge are associated
with the Saffir-Simpson Category 1–4 hurricanes. The SLOSH
model, a forecast model for hurricane-induced water levels for
the Gulf and Atlantic coasts (Jelesnianski, Chen, and Shaffer,
1992), was used in conjunction with the Applied Science
Associates, Inc. (ASA), Inundation Toolbox-Interpolation Tool
(Isaji and Knee, 2009). The SLOSH model displays the
expected surge heights along coasts. Surge heights are not
uniform along the coastline but depend on the hurricane track,
wind speed, and topography and bathymetry at the point where
the storm makes landfall (FEMA, 2003). Storm surge heights
were derived for the Providence, Rhode Island/Boston, Massa-
chusetts, and Ocean City, Maryland, storm basins in SLOSH
and were interpolated to a raster surface of the same extent
and resolution as the DEM, with the ASA Toolbox. Elevations
from the DEM were compared with the elevations from the
storm surge surfaces and probabilities of inundation at sentinel
sites were calculated using the z-score uncertainty technique
(described below). The uncertainty technique incorporated
errors associated with the elevation data sets and known
sources of error that are unique to the water surface modeled by
SLOSH. A site was considered at risk from inundation if it had
an elevation less than or equal to the water surface elevation
that was predicted for that location. To account for topographic
Journal of Coastal Research, Vol. 00, No. 0, 0000
0 Murdukhayeva et al.
features that might prevent inundation of inland areas, only
raster cells adjacent to the ocean or contiguous with other
inundated cells were included in the risk maps.
ArcGIS 10 software was used for all geospatial data
processing (ESRI, 2011).
Statistical ProceduresProbabilities of inundation for sentinel sites were calculated
for each of the sea level rise scenarios (0.6 m, 1 m, and 2 m) and
storm surge scenarios (Category 1–4 hurricanes). The land
surface elevation and modeled water surface were compared,
and probabilities of inundation at sentinel sites were calculated
using the z-score inundation uncertainty technique described
by NOAA Coastal Services Center (2010a). The standard
normal cumulative-distribution function was used to calculate
probabilities of inundation and the certainty of the prediction
given errors associated with the data and models (Ott and
Longnecker, 2010). Standard scores, or z-scores, were calcu-
lated at each sentinel site using the formula:
z�score ¼ Inundation level� Elevation
RMSETotal; ð1Þ
where RMSETotal is calculated as follows:
RMSETotal ¼ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiðRMSEElevationÞ2 þ ðRMSEWaterSurfaceÞ2
q: ð2Þ
The RMSE for LIDAR DEMs is calculated as follows:
RMSEElevation ¼
ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffii ¼ 1nðxLIDAR;i � xGPS;iÞ2
n
s; ð3Þ
where xLIDAR is the elevation from LIDAR at a single location,
and xGPS is the elevation as determined by GPS in the same
location. The RMSEs for the GPS survey elevations were
reported by Trimble Software. The RMSEs for water surfaces
were reported by VDatum. An area was considered at risk from
inundation if it had an elevation less than or equal to the water
surface elevation predicted for that location and was contigu-
ous with other inundated pixels. Because high topographic
Figure 2. Sea level rise scenarios at Cape Cod, Massachusetts. The blue areas represent areas at risk from inundation in 1-m and 2-m sea level rise scenarios. The
areas in orange represent areas whose elevations are under the 1-m and 2-m water surfaces but are unconnected to the ocean or other areas expected to be
inundated. (Color for this figure is available in the online version of this paper.)
Journal of Coastal Research, Vol. 00, No. 0, 0000
Sea Level Rise and Storm Surge in Northeastern National Parks 0
features might prevent inundation of low-elevation areas
further inland, we retained inundated locations that were
contiguous with flooded areas but did not retain inland sites
that were low elevation but were separated from the coast by a
landscape feature higher than the inundation elevation (see
Figure 2).
Descriptive statistics were calculated for each risk estimate
variable (e.g. probability of inundation with 1 m sea level rise,
inundation depth at 2 m sea level rise), and the variables were
tested for normality using the Shapiro-Wilk normality test.
Because data were usually nonnormally distributed, we
performed pairwise comparisons using the nonparametric
Wilcoxon signed rank test, and where there were three or
more groups in a comparison, we used the Kruskal-Wallis rank
sum test. A principal components analysis (PCA) was used to
reduce the large number of risk measures to a smaller number
of variables to develop a composite measure of inundation risk
at sentinel sites. All analyses were conducted using the
statistical software package R (R Development Core Team,
2011).
RESULTS
Inundation MappingEstimates of the extent of sea level rise using the bathtub
model are shown for selected parts of CACO (Figure 2) and
ASIS (Figure 3). Expected storm surge heights from Category
1–4 storms were modeled for CACO and ASIS (Table 2). Figure
4 demonstrates how differences in expected surge heights
translate to inundation-zone mapping in the northern extent of
the CACO. Total areas at risk from sea level rise and storm
surge inundation scenarios are presented in Table 3.
Elevations of Sentinel SitesThe elevations of sentinel sites that were obtained from RTK
GPS and those from LIDAR-derived DEMs were significantly
different (Table 4). At CACO, the LIDAR elevations were lower
than RTK GPS elevations were and had a mean difference of
1.19 6 0.10 m (mean 6 1 standard error [SE]). At ASIS, the
LIDAR elevations were often higher than RTK GPS elevations
were and had a mean (6 SE) difference of 0.29 6 0.04 m. For all
of our probability calculations, we used the RTK GPS–derived
elevations for sentinel sites.
Probabilities of InundationMean probabilities of inundation for each park’s set of
sentinel sites are reported in Tables 5 and 6. The complete list
of sentinel sites and their individual probabilities of inundation
can be found in Murdukhayeva (2012). As expected, inundation
risk increased as sea level rise height and storm intensity
increased. The mean probabilities of inundation at CACO were
significantly different for the two sea level rise scenarios
(Wilcoxon signed rank, V¼ 0, p , 0.001). Similarly, for ASIS,
the mean probabilities of inundation are significantly different
for the three sea level rise scenarios (H¼66.02, df¼2, p ,
0.0001). The mean probabilities of inundation for the four
storm surge scenarios are significantly different for both CACO
(H¼44.51, df¼3, p , 0.0001) and ASIS (H¼96.52, df¼3, p ,
0.0001).
Storm surge inundation risk increased dramatically with
increasing Category storms. Sentinel sites at Provincetown
Airport, Pleasant Bay Marsh, and Wellfleet Harbor,
Figure 3. Sea level rise scenarios near Verrazano Bridge, Assateague Island: (A) 0.6 m, (B) 1 m, and (C) 2 m. (Color for this figure is available in the online version
of this paper.)
Table 2. Storm surge heights (above MHHW) predicted by SLOSH.
Storm Class
Surge Height (m)
CACO ASIS
Category 1 0.34–1.52 0.43–1.77
Category 2 0.91–3.20 0.73–3.02
Category 3 1.34–5.33 2.56–4.30
Category 4 1.80–6.07 4.15–5.55
Journal of Coastal Research, Vol. 00, No. 0, 0000
0 Murdukhayeva et al.
Massachusetts, are at risk under Category 3 and 4
hurricane scenarios. Along the entire length of Assateague
Island, sentinel sites are at high risk. Under the Category 2
storm surge scenario, 22 sentinel sites have .75% proba-
bility of inundation, and under the Category 3 and 4
scenarios, all 34 sentinel sites have .75% chance of
inundation (Table 6).
Habitat Changes Predicted by SLAMM
The SLAMM model predicted habitat classes under
selected scenarios of sea level rise by 2100. Change matrices
(Tables 7 and 8) show the number of sentinel sites in each
habitat class initially and their expected conversions because
of sea level rise. At CACO, only two scenarios were modeled
because the vertical accuracy of the LIDAR data (the primary
Figure 4. Inundation expected from storm surges at Cape Cod, Massachusetts: (A) Category 1, (B) Category 2, (C) Category 3, and (D) Category 4. (Color for this
figure is available in the online version of this paper.)
Journal of Coastal Research, Vol. 00, No. 0, 0000
Sea Level Rise and Storm Surge in Northeastern National Parks 0
input driving the model) was inadequate to support the 0.6-m
scenario. In the Undeveloped dry land category, 10 sentinel
sites experienced conversions to the Transitional marsh
category, one sentinel site converted to Estuarine beach, and
one sentinel site converted to Open ocean after 2 m of sea level
rise (Table 7). In ASIS, many sentinel sites were in the
Undeveloped dry land class (24 of 34 total sites [71%]) and
experienced conversions to Transitional marsh (n¼ 9; 38%),
Estuarine beach (n ¼ 1; 4%), Ocean beach (n ¼ 3; 13%), and
Open ocean (n ¼ 11; 46%) with 2 m of sea level rise. Four
sentinel sites that started in the Irregularly flooded marsh
category converted to Salt marsh after 2 m of sea level rise
(Table 8).
Following the recommendation of Scarborough (2009),
similar land cover classes were aggregated for ease of
interpretability. The 11 possible SLAMM categories were
Table 3. Extent of inundation (in hectares) within each study area under bathtub model sea level rise scenarios and SLOSH hurricane scenarios.
Study Area
Sea Level Rise Scenario (m) Hurricane Class
0.6 1 2 Category 1 Category 2 Category 3 Category 4
CACO — 4423 5378 2720 3508 4234 5012
ASIS 4541 6224 8211 2607 6693 8147 8332
Table 4. Summary of sentinel site elevations. Differences between RTK GPS and LIDAR elevations were computed on a pairwise basis at each sentinel site.
Standard error values account for variability of elevations among sentinel sites and do not account for measurement errors. Test of the hypothesis that the
mean difference ¼ 0 was done using a Paired Sample Wilcoxon Test.
Study Area (n)
RTK GPS
(mean 6 SE m)
LIDAR
(mean 6 SE m)
Absolute Difference
(pairwise mean 6 SE m)
Test of Mean Pairwise Differences
V p
CACO (63) 13.20 6 1.45 12.02 6 1.46 1.19 6 0.10 1999 ,0.0001
ASIS (34) 1.39 6 0.07 1.65 6 0.07 0.29 6 0.04 39 ,0.0001
Table 5. Mean (6SE) probabilities of inundation at sentinel sites in each bathtub modeling scenario. The number of sites where the probability of inundation
exceeds 0.75 is shown in parentheses.
Study Area (n)
Inundation Scenario
0.6 m 1 m 2 m
CACO (63) — 0.064 6 0.029 (3) 0.179 6 0.047 (11)
ASIS (34) 0.130 6 0.51(3) 0.485 6 0.067 (11) 0.950 6 0.035 (32)
Table 6. Mean (6SE) probabilities of inundation at sentinel sites using the SLOSH storm scenarios. The number of sites where the probability of inundation
exceeds 0.75 is shown in parentheses.
Study Area (n)
Hurricane Class
Category 1 Category 2 Category 3 Category 4
CACO (63) 0.000 6 0.000 (0) 0.013 6 0.012 (1) 0.051 6 0.022 (1) 0.114 6 0.034 (5)
ASIS (34) 0.162 6 0.047 (2) 0.700 6 0.066 (22) 0.980 6 0.009 (34) 0.999 6 0.001 (34)
Table 7. CACO SLAMM conversion matrix. Numbers in parentheses indicate sentinel sites converted to each class in each inundation scenario (1 m and 2 m).
Initial Category
(No. of sentinel sites)
Predicted Category (1 m, 2 m sea level scenarios)
Dry Land
Developed
Dry Land
Undeveloped
Nontidal
Swamp
Transitional
Marsh
Salt
Marsh
Estuarine
Beach
Rocky
Intertidal
Irregularly
Flooded Marsh Open Ocean
Dry land developed (12) (12, 12)
Dry land undeveloped (46) (40, 34) (5, 10) (1, 1) (0, 1)
Nontidal swamp (1) (1, 0) (0, 1)
Transitional marsh (0)
Salt marsh (0)
Estuarine beach (0)
Rocky intertidal (2) (1, 0) (1, 2)
Irregularly flooded marsh (2) (1, 2) (1, 0)
Open ocean (0)
Journal of Coastal Research, Vol. 00, No. 0, 0000
0 Murdukhayeva et al.
Table 8. ASIS SLAMM conversion matrix. Numbers in parentheses indicate sentinel sites converted to each class in each inundation scenario (0.6 m, 1 m, 2
m).
Initial Category
(No. of sentinel sites)
Predicted Category (0.6 m, 1 m, 2 m sea-level scenarios)
Dry Land
Developed
Dry Land
Undeveloped
Inland
Fresh Marsh
Transitional
Marsh
Salt
Marsh
Estuarine
Beach
Tidal
Flat
Ocean
Beach
Irregularly
Flooded Marsh
Estuarine
Open Water
Open
Ocean
Dry land developed (1) (1, 1, 1)
Dry land undeveloped (24) (8, 2, 0) (7, 7, 9) (0, 1, 1) (8, 11, 3) (1, 3, 11)
Inland fresh marsh (1) (1, 1, 0) (0, 0, 1)
Transitional marsh (1) (1, 1, 0) (0, 0, 1)
Salt marsh (1) (1, 1, 0) (0, 0, 1)
Estuarine beach (1) (1, 1, 0) (0, 0, 1)
Tidal flat (0)
Ocean beach (1) (1, 1, 1)
Irregularly flooded marsh (4) (0, 1, 4) (4, 3, 0)
Estuarine open water (0)
Open ocean (0)
Figure 5. The SLAMM initial conditions and model output at Assateague Island. Eleven possible land-use classes in the study area (top panels), and land-use
classes aggregated into five broad groups (bottom panels). (Color for this figure is available in the online version of this paper.)
Journal of Coastal Research, Vol. 00, No. 0, 0000
Sea Level Rise and Storm Surge in Northeastern National Parks 0
reclassified into five land-cover categories for mapping appli-
cations: Upland, Forested wetland, Marsh, Beach, and Open
water. Figure 5 shows initial conditions and a 1-m sea level rise
output with the original classification scheme (top panels) and
with the aggregated classification scheme (bottom panels).
Figure 6 shows an area east of Calfpen Bay, Virginia (ASIS),
which showed great changes over the modeled scenarios; there
was a large increase in Marsh, Transitional marsh, and Open
Figure 6. The SLAMM output at Assateague Island. Eleven possible land-use classes in this study area aggregated into five broad groups for ease of
interpretation. (Color for this figure is available in the online version of this paper.)
Journal of Coastal Research, Vol. 00, No. 0, 0000
0 Murdukhayeva et al.
water areas. This area showed some of the greatest changes in
the two parks studied.
Maps with park-scale views of initial land-cover classes and
model-output land-cover classes for each sea level rise scenario
are found in Murdukhayeva et al. (2012).
Overall Inundation IndexThe three models yielded several measures of inundation
risk. We used PCA to reduce the large set of correlated risk
variables to a set of uncorrelated variables or principal
components (Table 9). The first principal component (PC1)
explained 63% of the total variation in risk measures at CACO
and 58% of the total variation in risk measures at ASIS. In all
cases, inundation probabilities had a negative loading, and
RTK GPS elevations had a positive loading on PC1 (Table 9).
Nearly all variables had similar loading values on PC1, thus
PC1 represents a ‘‘size’’ effect (August, 1983) and can be used
as an index of the overall risk of inundation. The other
principal components had high loadings on only one or two
variables and reflected specific risk factors. Furthermore, they
did not predict overall variation well and, therefore, were not
candidates for an overall risk index. The range of PC1 values
for each park was separated into five quintiles. Large, positive
PC1 scores for sentinel sites indicated that inundation was very
unlikely; large, negative PC1 scores indicated that inundation
was very likely (Table 10). Overall inundation index classifi-
cation at sentinel sites was mapped for CACO and ASIS (Figure
7).
DISCUSSION
Quality of the LIDAR DataThe best LIDAR data that were available to us were not as
accurate as expected and this affected how they could be used
in the models. The metadata for each of the two LIDAR-derived
DEMs reported a 15 cm vertical RMSE. To validate these
accuracy estimates, we calculated the vertical RMSE using
ground control points of high quality and accuracy (,2 cm
vertical and horizontal accuracy) collected using survey-grade
GPS. Our estimates of the RMSE of the LIDAR data were two
to three times larger than the RMSE values reported in the
metadata. This result had many implications in our study. We
were unable to reliably model the 0.6-m sea level rise scenario
at CACO based on the uncertainly of the LIDAR DEM used in
the modeling (NOAA, 2010). These results suggest that it is
prudent to independently test the accuracy of DEMs used in
inundation modeling.
Quality of the RTK GPS DataThe high accuracy of the elevations measured with RTK GPS
was an important asset for this project. The GPS field surveys
allowed us to collect accurate elevations of sentinel sites that
were hidden from LIDAR signals and sensors (NOAA CSC,
2010b). For example, we surveyed elevations of groundwater
monitoring wells located in forests with heavy canopy cover.
Furthermore, because these elevations were measured with
great accuracy using the RTK GPS (RMSE values of 0.6–8.7
cm), inundation probabilities could be calculated with greater
certainty. The application of RTK GPS is a promising solution
to elevation uncertainty issues in sea level rise inundation risk
assessments. It is important to note, however, that RTK GPS
protocols require operating a GPS base station at a location
that has been surveyed to within a few millimeters. For this
reason, a network of accurate geodetic control sites within 5 km
is a requirement for RTK GPS measurement at sentinel sites
(Murdukhayeva 2012).
The use of high accuracy RTK GPS data helped minimize the
error in estimating inundation risk. Because the error
associated with each sentinel site elevation was low, there
Table 9. Principal components analysis of risk variables. Class loadings for the first three principal components are provided for each variable. The GPS
elevation refers to site elevation as surveyed by GPS; C1_depth to C4_depth refer to the depth of inundation from Category 1–4 storms; Prob_60cm, Prob_1m,
and Prob_2m refer to probabilities of inundation given sea level rise scenarios of 60 cm, 1 m, and 2 m; SLAMM_60cm, SLAMM_1m, and SLAMM_2m refer to
conversion predictions (yes or no) by SLAMM given sea level rise scenarios of 60 cm, 1 m, and 2 m.
Variable
CACO ASIS
PC1 PC2 PC3 PC1 PC2 PC3
GPS elevation 0.37 0.23 0.09 0.42 0.09 �0.11
C1_depth �0.37 �0.23 �0.09 �0.39 �0.04 0.10
C2_depth �0.37 �0.23 �0.09 �0.33 0.46 �0.04
C3_depth �0.37 �0.23 �0.08 �0.40 0.20 �0.10
C4_depth �0.37 �0.23 �0.08 �0.37 �0.18 �0.06
Prob_60cm �0.19 0.47 �0.53 �0.28 �0.36 �0.45
Prob_1m �0.23 0.49 �0.34 �0.33 �0.25 �0.20
Prob_2m �0.27 0.30 0.41 �0.28 0.17 0.74
Slamm_60cm — — — �0.04 �0.70 0.42
Slamm_1m �0.25 0.39 0.23 — — —
Slamm_2m �0.29 0.17 0.59 — — —
% Variation explained 62.7 23.5 7.3 57.8 16.4 9.8
Cumulative variation (%) 62.7 86.3 93.6 57.8 74.2 84.0
Table 10. Inundation index class and corresponding PC1 raw scores. The
PC score classes were based on quintile categories of the data.
Inundation Index Class CACO ASIS
Very likely �5.54 to �1.87 �5.55 to �1.74
Likely �1.87 to �0.56 �1.74 to �0.43
As likely as not �0.56 to 0.60 �0.43 to 0.45
Unlikely 0.60 to 1.62 0.45 to 1.66
Very unlikely 1.62 to 5.56 1.66 to 6.00
Journal of Coastal Research, Vol. 00, No. 0, 0000
Sea Level Rise and Storm Surge in Northeastern National Parks 0
was less uncertainty associated with each modeled scenario.
Many of the sentinel sites had probabilities of inundation of
either 0 (very unlikely) or 1 (very likely) when the RTK GPS
elevation values for the sentinel sites were used. Murdukhaye-
va et al. (2012) found that this was not the case with
probabilities of inundation calculated using elevations at
sentinel sites based on LIDAR data, rather than the more
accurate RTK GPS; many probabilities of inundation were in
the range of 0.25 to 0.75 (rather than 0 or 1). Using higher-
accuracy elevation data for sentinel sites resulted in greater
certainty regarding the likelihood of inundation.
Habitat-Change Modeling
The SLAMM maps with aggregated land-cover classes can
provide useful information for assessing habitat transition
under various sea level rise scenarios. In presenting the
SLAMM output in maps, however, it is important to recognize
the uncertainty of habitat predictions because of uncertainty in
input elevation and land cover. The SLAMM model is
constrained by the quality of the data driving it (e.g. LIDAR-
derived DEMs). It was not possible to integrate high-accuracy
RTK GPS elevations into the model to enhance the results.
Figure 7. Overall inundation index for sentinel sites at CACO (left) and ASIS (right). Blue points indicate sites where inundation is very unlikely, and red points
indicate sites where inundation is very likely. (Color for this figure is available in the online version of this paper.)
Journal of Coastal Research, Vol. 00, No. 0, 0000
0 Murdukhayeva et al.
Another important baseline data set for the SLAMM model is
initial land-use classification. In this study, initial land-use
conditions were obtained, in part, from National Wetlands
Inventory maps created using aerial photography from 1988
(ASIS) and 1993 (CACO). Those maps did not reflect changes
during the past two decades. Furthermore, there is no way to
quantitatively determine the uncertainty associated with a
SLAMM prediction. These factors and others limit the utility of
the output results (Kirwan and Guntenspergen, 2009; Scar-
borough, 2009), particularly for park managers developing
mitigation strategies. Higher-resolution elevation data and
current land use data sets are becoming more widely available
and should address these deficits.
For the risk assessment, we were first interested in
predicting changes from land to open water (i.e. inundation).
The SLAMM results indicated that CACO had one site (1-m
scenario) and three sites (2-m scenario) convert to open water,
whereas, at ASIS, four sites (1-m scenario) and 12 sites (2-m
scenario) are predicted to convert to open water. The bathtub
model predicts greater amounts of inundation (Table 6). At
CACO, three (1-m scenario) and 11 sites (2-m scenario) have
probabilities of inundation .75%, and at ASIS, 11 (1-m
scenario) and 32 (2-m scenario) sites have probabilities of
inundation .75%. This difference in results supports the
argument for bathtub models over estimate inundation
(Mcleod et al., 2010; NOAA, 2010), but it may also indicate
that SLAMM underestimated inundation. An important
finding of our study was that the use of multiple methods
provided us a robust range of inundation possibilities across
plausible sea level scenarios.
Inundation ModelingBathtub modeling is a relatively straightforward technique
that can be used with off-the-shelf geospatial data, but the
resulting inundation surfaces can be inaccurate (Mcleod et al.,
2010; Poulter and Halpin, 2008). Typically, that inaccuracy
derives from subtle land barriers that are not discerned in the
DEM and can impede inland water flow. Similarly, structures
that permit water flow under the topographic surface, such as
culverts below roads, can contribute to error in modeled
inundation surfaces. Future human modifications of the
landscape as adaptation measures to sea level rise are not
included in simple bathtub models nor are changes to coastal
geophysical processes because of sea level rise, for example
breeching of barrier beaches that would significantly change
tidal dynamics of lagoon systems behind the barrier (Anthony
et al., 2009). By using high-quality LIDAR data for our DEMs
and by excluding pixels that were not connected to the coast,
our bathtub model results can serve as a preliminary
assessment of potential inundation regions in these study
areas.
One of the most valuable products of this assessment is the
range of storm surge heights modeled by SLOSH (Table 4).
Those surge heights can be used for predictions of hurricane
impacts in the near future. For example, the SLOSH model
predicted extensive inundation in Provincetown, Massachu-
setts, and the salt marsh sites in Cape Cod, Massachusetts. The
potential damage to a sentinel-site resource under conditions of
storm surge inundation varies depending on the nature of the
site. Buildings or historical artifacts could face extensive
damage, whereas certain habitats or hard infrastructure
(roads, geodetic monuments) might not be damaged at all
during a brief period of inundation. Therefore, it is important
that the NPS evaluate risk for each site and develop mitigation
plans accordingly.
Implications for Sentinel SitesFor the most part, Cape Cod’s sentinel sites were located
inland, on relatively high elevations; 39 of the 63 sentinel sites
(62%) had elevations .5 m above the MHHW, and, as a result,
many sites are not at risk from sea level rise or storm surge
inundation. For example, the base of Highland Lighthouse
(Truro, Massachusetts) is located at 39.34 m NAVD88 or 38.41
m above MHHW (Figure 8). Although we understand that
inundation modeling does not identify those sites .5 m above
the MHHW as at risk, they could face other impacts of
changing coastal patterns (erosion, etc.). Several sites that
were found to be at risk are very close to current tide levels. For
example, one sentinel site at CACO, a culvert near Wellfleet
Harbor, was almost submerged at high tide on our survey trip
(Figure 8). One polyvinyl chloride (PVC) survey marker in
Pleasant Bay and National Geodetic Survey (NGS) monuments
at the Provincetown airport are at risk as well. These findings
agree closely with the results of the Hammar-Klose et al. (2003)
Coastal Vulnerability Index assessment. In their study, 1.5-km
segments of shoreline were ranked as being of low to high
vulnerability, using an index that combined geological and
physical variables. The high-vulnerability sentinel sites, as
ranked by the composite Inundation Index we derived from the
PCA, tend to appear immediately inland of those shoreline
segments with a high CVI vulnerability rank.
At Assateague Island, all of the sentinel sites were at low
elevations ,2.6 m. Many of them are at risk from sea level rise
and storm surge from large hurricanes. The CVI assessment
and spatial pattern of the overall Inundation Index at ASIS
correspond with each other (Pendleton, Williams, and Thieler,
2004). Disagreement occurs at points near the Chincoteague
Inlet, Virginia. The CVI rated that shoreline as low vulnera-
bility because of high accretion rates, and the Inundation Index
rated sentinel sites in that area as high vulnerability because of
the low elevations.
It is inappropriate to compare the frequency of inundation at
sentinel sites between the national parks in our study because
resource managers at CACO and ASIS chose very different
kinds of sites to serve as sentinel sites. The CACO sites
consisted of a variety of historical and natural resources, as
well as important infrastructure. The ASIS sentinel sites were
monumented reference locations at low elevations. If the CACO
resource managers had, like those at ASIS, selected only
sentinel sites used to monitor sea level rise then those two sets
of sentinel sites would have had more comparable landscape
placement, and the frequency of inundation could have been
quite similar despite the very different geomorphology of the
two parks.
Conclusions and Directions for Future WorkThis study assessed inundation risk at 97 sentinel sites
located in two northeastern U.S. coastal national parks.
Rigorous revalidation of the LIDAR-derived data inputs was
Journal of Coastal Research, Vol. 00, No. 0, 0000
Sea Level Rise and Storm Surge in Northeastern National Parks 0
essential to making appropriate use of the models. The
methodology we used can be applied at other coastal parks
and can be repeated at the same parks with improved data,
alternate scenarios, or additional sentinel sites. We used three
different modeling approaches because we were interested in
inundation and habitat changes under various sea level rise
scenarios as well as inundation from storm surges. Each model
provided unique information that served our management
needs. Depending on the resource issues and management
needs of other coastal sites, all three models may or may not be
required.
Our choice of models was driven by the goal to make our
methods transportable to other coastal settings and imple-
mentable using off-the-shelf data. Accurate elevation data,
such as that provided by LIDAR technology, are an essential
requisite for inundation modeling. These analyses could not be
performed with coarse data, such as the NED. All the models
suffer from the naı̈ve assumption that coastal geomorphology
does not change as sea level rises or storm surge overwashes
inland features. As coastal barrier/lagoon systems breach or
dune systems wash away, inundation dynamics will change,
and those changes are not accounted for in the models we have
chosen.
Models of sea level rise and storm surge are continually being
refined, and inundation probabilities can be recomputed as new
models are developed. Scenarios for sea level rise are evolving
as new data from satellite altimetry and ice-melt studies are
acquired. The U.S. National Weather Service updates SLOSH’s
storm-surge predictions for regional basins following large
hurricane events. The LIDAR data are being acquired for large
regions of the United States coast to evaluate sea level rise and
storm-surge risk. For example, the USGS recently completed
Figure 8. (Top) Sentinel site with low relative likelihood of inundation. Highland Lighthouse in North Truro, Massachusetts. (Bottom) Low-lying sentinel site at
Cape Cod National Seashore. A culvert near a bike path in Wellfleet, Massachusetts. (Color for this figure is available in the online version of this paper.)
Journal of Coastal Research, Vol. 00, No. 0, 0000
0 Murdukhayeva et al.
an extensive LIDAR acquisition for coastal counties of the
northeast United States. The use of RTK GPS technology to
collect accurate elevations at sentinel sites is a promising
undertaking that will improve predictions of inundation risk
based on different possible scenarios and models. The growing
network of stable, permanent, and accurate geodetic control
sites to form a backbone of base station locations makes it
possible to greatly expand the areas where rapid collection of
high-resolution elevation data with RTK GPS are possible
(Murdukhayeva et al., 2012).
POSTSCRIPTWe visited 11 sentinel sites at ASIS immediately following
Hurricane Sandy to determine the accuracy of our inundation
modeling. Peak surge was 29 October 2012 at 1730. The Ocean
City, Maryland, tide gauge (1.5 km from the closest sentinel
site) recorded a surge elevation of 1.4 m NAVD88. We inspected
each sentinel site for signs of inundation, such as sand scouring
and evidence of wrack. Seven sites were at elevations above 1.4
m, and of those, five (71%) were not inundated. The two that did
flood were very close to the surge elevation (1.5 m and 1.7 m).
No sentinel sites above 1.9 m were inundated. Four sites were
below the surge elevation, and of these, three (75%) were
flooded. The site that did not flood was within a meter or two
(horizontally) of the wrack line. The mean (6SE) PC1
inundation index score for flooded sentinel sites was 0.59 6
0.48 (n¼5), and the mean (6SE) PC1 score for sites that did not
flood was 2.97 6 0.94 (n ¼ 6). The flooded sites had a
significantly lower inundation index (i.e. higher risk of
inundation) than did nonflooded sites (Wilcoxon two-sample
test, V¼ 26, p¼ 0.025).
ACKNOWLEDGMENTSThis research was supported by NPS Task Agreement
J4531090800—Assessing Inundation Risk From Sea Level
Rise and Storm Surge in Coastal National Parks Using High
Accuracy Geodetic Control—administered by the North
Atlantic Coast Cooperative Ecosystem Studies Unit at the
University of Rhode Island Coastal Institute. We gratefully
acknowledge the technical and logistic support we received
from the Northeast Coastal and Barrier Network of the NPS
Inventory and Monitoring Program. The project has benefited
from the contributions and ideas of Tim Smith, Rob Thieler,
Donald Cahoon, Kelly Knee (Applied Science Associates), Neil
Winn, Marc Albert, Mark Adams, Charles Roman, YQ Wang,
Chris Damon, Roland Duhaime, Sarah Stevens, Dennis
Skidds, Galen Scott, Tiffany-Lane Davis, Heather Grybas,
Doug Marcy, and many others. The thoughtful comments by
three anonymous reviewers significantly improved the clarity
of the paper.
LITERATURE CITEDAnthony, A.; Atwood, J.; August, P.; Byron, C.; Cobb, S.; Foster, C.;
Fry, C.; Gold, A.; Hagos, K.; Heffner, L.; Kellogg, Q.; Lellis, K.;Opaluch, J.; Oviatt, C.; Pfeiffer-Herbert, A.; Rohr, N.; Smith, L.;Smythe, T.; Swift, J., and Vinhateiro, N., 2009. Coastal lagoons andclimate change: Ecological and social ramifications in temperateecosystems. Ecology and Society, 14(1), 8–19.
Ashton, A.D.; Donnelly, J.P., and Evans, R.L., 2008. A discussion ofthe potential impacts of climate change on the shorelines of the
northeastern USA. Mitigation and Adaptation Strategies forGlobal Change, 13(7), 719–743.
August, P.V., 1983. The effects of habitat complexity and heteroge-neity in structuring tropical mammal communities. Ecology, 64(6),1495–1507.
Bender, M.A.; Knutson, T.R.; Tuleya, R.E.; Sirutis, J.J.; Vecchi, G.A.;Garner, S.T., and Held, I.M., 2010. Modeled impact of anthropo-genic warming on the frequency of intense Atlantic hurricanes.Science, 327(5964), 454–458.
Bonisteel, J.M.; Nayegandi, A.; Wright, C.W.; Brock, J.C., and Nagle,D.B., 2009. Experimental Advanced Airborne Research Lidar(EAARL) Data Processing Manual. Reston, Virginia: U.S. Geolog-ical Survey, Open File Report 2009-1078, 38p.
Bonisteel-Cormier, J.M.; Nayegandhi, A.; Brock, J.C.; Wright, C.W.;Nagle, D.B.; Klipp, E.S.; Vivekanandan, S.; Fredericks, X., andStevens, S., 2010. EAARL Coastal Topography and Imagery–Assateague Island National Seashore, Maryland and Virginia,post-Nor’Ida, 2009. Reston, Virginia: U.S. Geological Survey, DataSeries 559, 1 DVD.
Brock, J.C.; Wright, C.W.; Patterson, M.; Nayegandhi, A., andTravers, L.J., 2007. EAARL Topography–Cape Cod NationalSeashore, 2005. Reston, Virginia: U.S. Geological Survey, OpenFile Report 2007-1375, http://pubs.usgs.gov/of/2007/1375/start.html.
Brown, I., 2006. Modelling future landscape change on coastalfloodplains using a rule-based GIS. Environmental Modelling andSoftware, 21(10), 1479–1490.
Cazenave, A. and Nerem, R.S., 2004. Present-day sea level change:observations and causes. Review of Geophysics, 42(3), RG3001.doi:10.1029/2003RG000139.
Church, J.A. and White, N.J., 2006. A 20th century acceleration inglobal sea-level rise. Geophysical Research Letters, 33(1), L01602.doi:10.1029/2005GL024826.
Engelhart, S.E.; Peltier, W.R., and Horton, B.P., 2011. Holocenerelative sea-level changes and glacial isostatic adjustment of theU.S. Atlantic coast. Geology, 39(8), 751–754.
ESRI (Environmental Systems Research Institute), 2011. ArcGISDesktop: Release 10. Redlands, California: Environmental SystemsResearch Institute.
FEMA (Federal Emergency Management Agency), 2003. SLOSHDisplay Training. http://www.fema.gov/pdf/plan/prevent/nhp/slosh_display_training.pdf.
Gao, J., 2007. Towards accurate determination of surface height usingmodern geoinformatic methods: possibilities and limitations.Progress in Physical Geography, 31(6), 591–605.
Gesch, D.B., 2007. The National Elevation Dataset. In: Maune, D.(ed.), Digital Elevation Model Technologies and Applications: TheDEM Users Manual 2nd edition. Bethesda, Maryland: AmericanSociety for Photogrammetry and Remote Sensing, pp. 99–118.
Gesch, D.B., 2009. Analysis of lidar elevation data for improvedidentification and delineation of lands vulnerable to sea-level rise.In: Brock, J.C. and Purkis, S.J. (eds.), Coastal Applications ofAirborne LIDAR. Journal of Coastal Research, Special Issue, No.53, pp. 49–58.
Gregg, R.M., 2010. Planning for sea level rise in Rhode Island’sCoastal Management Program: case study on a project of the RhodeIsland Coastal Resources Management Council. http://www.cakex.org/case-studies/1546.
Gutierrez, B.T.; Williams, S.J., and Thieler, E.R., 2007. Potential forShoreline Changes Due to Sea-Level Rise along the U.S. Mid-Atlantic Region. Reston, Virginia: U.S. Geological Survey, Open-File Report 2007-1278, 30p.
Hammar-Klose, E.S.; Pendleton, E.A.; Thieler, E.R., and Williams,S.J., 2003. Coastal Vulnerability Assessment of Cape Cod NationalSeashore (CACO) to Sea-Level Rise. Reston, Virginia: U.S.Geological Survey, Open-File Report 02–233, 23p.
Harvey, N. and Nicholls, R., 2008. Global sea-level rise and coastalvulnerability. Sustainability Science, 3(1), 5–7.
Hennecke, W.G. and Cowell, P.J., 2000. GIS modeling of impacts of anaccelerated rate of sea-level rise on coastal inlets and deeplyembayed shorelines. Environmental Geosciences, 7(3), 137–148.
Journal of Coastal Research, Vol. 00, No. 0, 0000
Sea Level Rise and Storm Surge in Northeastern National Parks 0
Hu, A.; Meehl, G.A.; Han, W., and Yin, J., 2009. Transient response ofthe MOC and climate to potential melting of the Greenland IceSheet in the 21st century. Geophysical Research Letters, 36(10).doi:10.1029/2009GL037998.
IPCC (Intergovernmental Panel on Climate Change), 2007. ClimateChange 2007:The Physical Science Basis. Contribution of WorkingGroup I to the Fourth Assessment Report (AR4). New York:Cambridge University Press, 996p.
Irish, J.L.; Frey, A.E.; Rosati, J.D.; Olivera, F.; Dunkin, L.M.;Kaihatu, J.M.; Ferreira, C.M., and Edge, B.L., 2010. Potentialimplications of global warming and barrier island degradation onfuture hurricane inundation, property damages, and populationimpacted. Ocean and Coastal Management, 53(10), 645–657.
Isaji, T. and Knee, K., 2009. Interpolation Tool for Handling StormSurge Spatial Variability for Inundation Simulation. South Kings-town, Rhode Island: Applied Science.
Jarvinen, B.R. and Lawrence, M.B., 1985. An evaluation of theSLOSH storm surge model. Bulletin of the American MeteorologicalSociety, 66(11), 1408–1411.
Jelesnianski, C.P.; Chen, J., and Shaffer, W.A., 1992. SLOSH: Sea,Lake, and Overland Surges from Hurricanes. Silver Spring,Maryland: National Oceanic and Atmospheric Administration,Technical Report NWS 48, 71p.
Kirshen, P.; Watson, C.; Douglas, E.; Gontz, A.; Lee, J., and Tian, Y.,2008. Coastal flooding in the Northeastern United States due toclimate change. Mitigation and Adaptation Strategies for GlobalChange, 13(5), 437–451.
Kirwan, M.L. and Guntenspergen, G.R., 2009. Accelerated sea-levelrise—a response to Craft et al. Frontiers in Ecology and theEnvironment, 7(2), 126–127.
Lewis, M.; Horsburgh, K.; Bates, P., and Smith, R., 2011. Quantifyingthe uncertainty in future coastal flood risk estimates for the U.K.Journal of Coastal Research, 27(5), 870–881.
Lin, N.; Emanuel, K.A.; Smith, J.A., and Vanmarcke E., 2010. Riskassessment of hurricane storm surge for New York City. Journal ofGeophysical Research, 115: D18121. doi:10.1029/2009JD013630.
Marindin, H.L., 1891. On the changes in the shoreline and anchorageareas of Cape Cod (or Provincetown Harbor) as shown by acomparison of surveys made between 1835, 1867, and 1890. SilverSprings, Maryland: U.S. Coast and Geodetic Survey Report 1891, pt2, app. 8, pp. 283–288.
MassGIS, 2005. MassGIS LIDAR for Metropolitan Boston. Boston:Commonwealth of Massachusetts Information Technology Divi-sion, Office of Geographic Information.
McInnes, K.L.; Walsh, K.J.E.; Hubbert, G.D., and Beer, T., 2003.Impact of sea-level rise and storm surge on a coastal community.Natural Hazards, 30(2), 187–207.
Mcleod, E.; Poulter, B.; Hinkel, J.; Reyes, E., and Salm, R., 2010. Sea-level rise impact models and environmental conservation: a reviewof models and their applications. Ocean and Coastal Management,53(9), 507–517.
Murdukhayeva, A., 2012. Assessment of inundation risk from sealevel rise and storm surge in coastal National Parks. Kingston,Rhode Island: University of Rhode Island, Master’s Thesis, 118p.
Murdukhayeva, A.; Bradley, M.; Shaw, N.; LaBash, C.; Grybas, H.;Davis, T.; August, P.V.; Smith T., and Duhaime, R., 2012. Usinghigh accuracy geodesy to assess risk from climate change in coastalNational parks. In: Weber, S., ed. Rethinking Protected Areas in aChanging World: Proceedings of the 2011 George Wright SocietyBiennial Conference on Parks, Protected Areas, and Cultural Sites.Hancock, Michigan: The George Wright Society, pp. 252–259.
NOAA (National Oceanic and Atmospheric Administration), 2006.The Coastal Change Analysis Program (C-CAP) Regional LandCover. http://www.csc.noaa.gov/digitalcoast/data/ccapregional.
NOAA, 2007. National Oceanic and Atmospheric Administration,Center for Operational Oceanographic Products and Services: Tidesand Currents. http://tidesandcurrents.noaa.gov.
NOAA, 2010. Technical Considerations for Use of Geospatial Data inSea Level Change Mapping and Assessment. Silver Spring,
Maryland: National Oceanic and Atmospheric Administration,NOS 2010-01, 141p.
NOAA, 2011. VDatum. http://vdatum.noaa.gov.NOAA CSC (National Oceanic and Atmospheric Administration
Coastal Services Center), 2010a. Mapping Inundation Uncertainty.Charleston, South Carolina: NOAA Coastal Services Center, 10p.
NOAA CSC, 2010b. Lidar Data Collected in Marshes: Its Error andApplication for Sea Level Rise Modeling. Charleston, SouthCarolina: NOAA Coastal Services Center, 22p.
NPS (National Park Service), 2009. Cape Cod National Seashore ResourceBrief: Salt Marsh Elevation. http://www.nps.gov/caco/naturescience/upload/MMsaltmarshelevationfactsheet_corrected-2.pdf.
Ott, R.L. and Longnecker, M., 2010. An Introduction to StatisticalMethods and Data Analysis. Belmont, California: Brooks/Cole,1273p.
Overpeck, J.T. and Weiss, J.L., 2009. Projections of future sea levelrise becoming more dire. Proceedings of the National Academy ofSciences of the United States of America, 106(51), 21461–21462.
Pendleton, E.A.; Williams, S.J., and Thieler, E.R., 2004. CoastalVulnerability Assessment of Assateague Island National Seashore(ASIS) to Sea-Level Rise. Woods Hole, Massachusetts: U.S.Geological Survey, Open File Report 2004-1020, 20p.
Pfeffer, W.T.; Harper, J.T., and O’Neel, S., 2008. Kinematicconstraints on glacier contributions to 21st-century sea level rise.Science, 321(5894), 1340–1343.
Poulter, B. and Halpin, P.N., 2008. Raster modelling of coastalflooding from sea-level rise. International Journal of GeographicInformation Science, 22(2), 167–182.
R Development Core Team, 2011. R: A Language and Environmentfor Statistical Computing. Vienna, Austria: R Foundation forStatistical Computing, 409p
Rahmstorf, S., 2007. A semi-empirical approach to projecting futuresea-level rise. Science, 315(5810), 368–370.
Sallenger, A.H.; Doran, K.S., and Howd, P.A., 2012. Hotspot ofaccelerated sea-level rise on the Atlantic coast of North America.Nature Climate Change. 2(12), 884–888 [Letter]. doi:10.1038/nclimate1597.
Scarborough, R.W., 2009. Application of the Sea Level Rise AffectingMarsh Model (SLAMM) Using High Resolution Data at PrimeHook National Wildlife Refuge. Dover, Delaware: DelawareDepartment of Natural Resources and Environmental Control, 62p.
Scawthorn, C.; Blais, N.; Seligson, H.; Tate, E.; Mifflin, E.; Thomas,W.; Murphy, J., and Jones, C., 2006a. HAZUS-MH flood lossestimation methodology, I: overview and flood hazard character-ization. Natural Hazards Review, 7(2), 60–71.
Scawthorn, C.; Flores, P.; Blais, N.; Seligson, H.; Tate, E.; Chang, S.;Mifflin, E.; Thomas, W.; Murphy, J.; Jones, C., and Lawrence, M.,2006b. HAZUS-MH flood loss estimation methodology, II: damageand loss assessment. Natural Hazards Review, 7(2), 72–81.
Sto. Domingo, N.D.; Paludan, B.; Madsen, H.; Hansen, F., and Mark,O., 2010. Climate Change and Storm Surges: Assessing Impacts onYour Coastal City Through Mike Flood Modeling. Copenhagen,Denmark: DHI Water, Environment and Health Report, 11p.
Tamisiea, M.E. and Mitrovica, J.X., 2011. The moving boundaries ofsea level change: understanding the origins of geographic variabil-ity. Oceanography, 24(2), 24–39.
USFWS (U. S. Fish and Wildlife Service), 2010. National WetlandsInventory. http://www.fws.gov/wetlands.
Van Drie, R.; Milevski, P., and Simon, M., 2010. Assessment of sealevel rise and climate change impacts using ANUGA. 19th NSWCoastal Conference (Batemans Bay, Australia), 15p.
Vermeer, M. and Rahmstorf, S., 2009. Global sea level linked to globaltemperature. Proceedings of the National Academy of Sciences,106(51), 21527–21532. doi:10.1073/pnas.0907765106.
Warren Pinnacle Consulting, 2010. SLAMM: Sea Level AffectingMarshes Model. http://warrenpinnacle.com/prof/SLAMM.
Yin, J.; Schlesinger, M., and Stouffer, R., 2009. Model projections ofrapid sea level rise on the northeast coast of the United States.Nature Geoscience, 2(4), 262–266. doi:10.1038/ngeo462.
Journal of Coastal Research, Vol. 00, No. 0, 0000
0 Murdukhayeva et al.