Estuarine Responses to Long‐Term Changes in Inlets, Morphology, and Sea Level Rise

Abstract

Pamlico Sound, a large back‐barrier estuary in North Carolina, is under threat of climate change due to increased storm activity and sea level rise. The response of this system is investigated by considering what has already happened during changes in sea level over the late Holocene epoch. The hydrodynamic changes that occurred in response to geomorphic evolution are simulated using a 3‐D numerical model for four distinct “time‐slice” scenarios. To accomplish this, the present‐day bathymetry was obtained from a high‐resolution digital elevation model, and paleobathymetric grids were developed from sediment cores and seismic observations. Using the same hydrodynamic forcing for each geomorphic scenario, the models are compared to assess the combined response to: different inlets connecting the back‐barrier estuary to the ocean, changes in basin geomorphology due to sedimentation, and sea level rise. The results indicate that these factors have a considerable effect on hydrodynamics, waves, and salinity in the estuary. The time‐averaged tidal ranges were up to 3 times as high for the past environments in comparison with present‐day water level elevations, and maximum current velocities were over 3 times higher in regions close to paleo‐inlets. The simulations for each time slice suggest that the salinity distribution in Pamlico Sound is strongly influenced by the hydraulic connectivity with other estuaries and the number and size of tidal inlets through the barrier island system. The results indicate that changes to barrier systems induce strong, nonuniform, and complex responses in back‐barrier estuaries with regime shifts in hydrodynamic energy and water mass properties.

Full text

Estuarine Responses to Long‐Term Changes in Inlets,
Morphology, and Sea Level Rise
Ryan P. Mulligan
1
, David J. Mallinson
2
, Gregory J. Clunies
1
, Alexander Rey
1
, Stephen J. Culver
2
,
Nick Zaremba
3
, Eduardo Leorri
2
, and Siddhartha Mitra
2
1
Department of Civil Engineering, Queen's University, Kingston, Ontario, Canada,
2
Department of Geological Sciences,
East Carolina University, Greenville, NC, USA,
3
Department of Earth Sciences, Syracuse University, Syracuse, NY, USA
Abstract Pamlico Sound, a large back‐barrier estuary in North Carolina, is under threat of climate
change due to increased storm activity and sea level rise. The response of this system is investigated by
considering what has already happened during changes in sea level over the late Holocene epoch. The
hydrodynamic changes that occurred in response to geomorphic evolution are simulated using a 3‐D
numerical model for four distinct “time‐slice”scenarios. To accomplish this, the present‐day bathymetry
was obtained from a high‐resolution digital elevation model, and paleobathymetric grids were developed
from sediment cores and seismic observations. Using the same hydrodynamic forcing for each geomorphic
scenario, the models are compared to assess the combined response to: different inlets connecting the
back‐barrier estuary to the ocean, changes in basin geomorphology due to sedimentation, and sea level rise.
The results indicate that these factors have a considerable effect on hydrodynamics, waves, and salinity in
the estuary. The time‐averaged tidal ranges were up to 3 times as high for the past environments in
comparison with present‐day water level elevations, and maximum current velocities were over 3 times
higher in regions close to paleo‐inlets. The simulations for each time slice suggest that the salinity
distribution in Pamlico Sound is strongly influenced by the hydraulic connectivity with other estuaries and
the number and size of tidal inlets through the barrier island system. The results indicate that changes to
barrier systems induce strong, nonuniform, and complex responses in back‐barrier estuaries with regime
shifts in hydrodynamic energy and water mass properties.
1. Introduction
Climate change has significant implications for the coastal and marine ecosystems and associated socioeco-
nomic systems of coastal communities that depend on them (Harley et al., 2006). Low‐lying coastal zones,
which contain a disproportionally high percentage of the world's population in a relatively small land area
(McGranahan et al., 2007), are at risk from climate change and associated sea level rise (SLR). Rising sea
level can lead to the inundation, erosion, and migration of natural coastal landforms and affect the salinity
distributions in coastal bays and estuarine systems (Scavia et al., 2002). Low‐lying coastal areas are suscep-
tible to increasing frequency in barrier overwash and erosion due to SLR (Bernstein et al., 2007), in combi-
nation with projected increasing tropical cyclone frequency arising from climate change (Mann et al., 2009;
Smith, 1999). There is evidence of major geomorphological changes in the past along the U.S. East Coast, as
widespread barrier island segmentation coincides with a peak in Atlantic hurricane activity during the
Medieval Climate Anomaly (MCA) ~1,100 to 900 calibrated years before present (cal yr BP) (Culver et al.,
2007; Donnelly et al., 2015; Donnelly & Woodruff, 2007; Grand Pre et al., 2011; Mallinson et al., 2011;
Mann et al., 2009) and increased rate of SLR (González & Törnqvist, 2009; Kemp et al., 2009, 2011, 2017).
Understanding the impacts of climate change and SLR on these dynamic systems is critical to the manage-
ment of rapidly developing coastal communities.
Numerical models are tools that can be used to advance our understanding of the hydrodynamic conditions
that may likely have occurred corresponding to different geomorphic settings over long time scales of hun-
dreds to thousands of years. SLR has contributed to dramatic changes of the coastal landscape over the
Holocene and a review of hydrodynamic response of coastal systems to SLR is provided by Passeri et al.
(2015). To assess the impacts of SLR on storm surge flooding over a shorter time scale, Bilskie et al. (2014)
used the Advanced Circulation model at three modern‐day times (1960, 2005, and 2050) with alterations
to bathymetry, land use, and land cover of the coast to demonstrate the nonlinearity of coastal flooding
©2019. American Geophysical Union.
All Rights Reserved.
RESEARCH ARTICLE
10.1029/2018JC014732
Key Points:
•A modeling approach is developed
for simulating the response of an
estuary to different
geomorphological conditions over a
long time scale
•Geological observations were used
to define inlet locations, basin
morphology, and sea levels in the
hydrodynamic model
•The results indicate controls on the
water levels, current velocities, and
salinity distributions in
paleo‐estuarine environments
Correspondence to:
R. P. Mulligan,
ryan.mulligan@queensu.ca
Citation:
Mulligan, R. P., Mallinson, D. J.,
Clunies, G. J., Rey, A., Culver, S. J.,
Zaremba, N., et al. (2019). Estuarine
responses to long‐term changes in
inlets, morphology and sea level rise.
Journal of Geophysical Research:
Received 31 OCT 2018
Accepted 17 NOV 2019
Accepted article online 28 NOV 2019
MULLIGAN ET AL.
Oceans,124, https://
doi.org/10.1029/2018JC014732
9235–9257.
Published online 17 DEC 2019
9235

driven by SLR and hurricane‐driven storm surge. In the Wadden Sea, a large back‐barrier estuarine system,
the sediment transport patterns have been simulated using the Delft3D model in response to extensive
coastal defense work over the last 400 yr (Elias & van der Spek , 2006) and in response to future SLR over
the next 100 yr (Wang et al., 2018). These studies highlight the interdependency between barrier island tidal
inlets and the back‐barrier morphology. Moran et al. (2014) used the Delft3D model to investigate the geo-
logic evolution of Currituck Sound, a lagoon associated with the Albemarle‐Pamlico Estuarine System
(APES) in North Carolina, USA. The hydrodynamic model was used to simulate tidal flows at time intervals
spanning 4,000 yr using bathymetry developed from geologic observations, with results indicating that
strong tidal currents in Currituck Sound were constrained to the immediate surroundings of the inlets.
The North Carolina coast is dominated by ocean beaches along the Outer Banks barrier island chain that
protect the system of interconnected back‐barrier estuaries. Previous geological studies have focused on
the Holocene geologic evolution of the APES, and it has been shown that significant changes to the Outer
Banks barrier island chain have occurred in the past, based on geophysical and sediment core observations
(Culver et al., 2007; Grand Pre et al., 2011; Mallinson et el., 2010, 2011; Zaremba et al., 2016). A detailed ana-
lysis of marsh cores and SLR in the APES is given by Kemp et al. (2017). Morphological changes of this
coastal region began around 9,000 cal yr BP with the flooding of incised fluvial valleys due to rapidly rising
relative sea level (~5 mm/yr; Horton et al., 2009) and formation of a broad estuary protected by low‐lying
barrier islands (Mallinson et al., 2018; Zaremba et al., 2016). Grain sizes and foraminiferal assemblages in
sediment cores indicate that over time conditions in the system shifted regimes. The regime shifts are from
a low‐energy brackish‐salinity estuarine environment with extensive barrier island protection and few inlets
to a higher‐energy high salinity coastal marine environment and significant segmentation of the barrier
islands (Culver et al., 2007; Grand Pre et al., 2011). Increased hurricane frequency during the MCA likely
resulted in a temporary segmentation due to breaching of the southern Outer Banks barrier islands allowing
for increased marine influence in the back‐barrier estuarine environment (Culver et al., 2007). Over the past
500 yr, estuarine deposits continued to accumulate (Peek et al., 2014) coinciding with a lower rate of relative
SLR from ~500 to 100 cal yr BP (Kemp et al., 2009) and the reestablishment of the barrier islands and low‐
energy back‐barrier estuary of the present day. The present‐day salinity distributions throughout the APES
have been described through observations and modeling studies. Wells and Kim (1989) synthesized histor-
ical observations and prescribed climatological mean surface salinity maps for the APES during both spring
(April) and fall (October) seasons. Jia and Li (2012) used the Regional Ocean Modeling System to determine
circulation dynamics and salt balance within the APES, finding that salinity is controlled by the relative
strength of inputs of freshwater from the four major rivers in comparison to the saltwater inflows occurring
at inlets as a result of tidal pumping. Brown et al. (2014) used the Delft3D numericalmodel to simulate fresh-
water and dissolved organic carbon transport in the Neuse River Estuary, a subestuary of the APES, demon-
strating the importance of stratification in the Neuse River Estuary and indicating a well‐mixed water
column at the estuary mouth in Pamlico Sound.
The objective of this study is to understand the responses in Pamlico Sound to periods of widespread barrier
island segmentation that have occurred in the past, in order to develop knowledge of potential future
changes that may occur as barrier island continuity and estuary morphology respond to SLR and storms.
To achieve this, paleobathymetric data presented in Zaremba et al. (2016) were used with varying barrier
island/inlet configurations developed from sedimentary and geophysical data (Culver et al., 2006 , 2007;
Grand Pre et al., 2011; Mallinson et al., 2011). These data are used to define four distinct geomorphic condi-
tions that occurred during the evolution of this coastal system and coupled numerical models are used to
numerically simulate the hydrodynamics and waves corresponding to each scenario. This investigation
demonstrates the significance of geomorphic changes to physical processes that broadly affect erosion, sedi-
ment transport and accumulation, as well as ecosystems via changes from estuarine to marine salinity con-
ditions. The application of this method is important for a more thorough interpretation of the geological
record in coastal systems, as well as for understanding the range of possible future changes.
2. Observations
Present‐day observations of winds, water levels, currents, river discharge, salinity, and sediment character-
istics were collected at various locations throughout Pamlico Sound (Figure 1a). Geophysical and core‐
10.1029/2018JC014732
Journal of Geophysical Research: Oceans
MULLIGAN ET AL. 9236

derived stratigraphic, lithologic, and micropaleontological observations (Figure 1b) were used to produce
paleobathymetric maps. These maps represent the estuarine geomorphology at four distinct times, herein
referred to as geomorphic “time‐slice”scenarios, over the last 4,000 yr (Culver et al., 2007; Foley, 2007;
Metger, 2009; Grand Pre et al., 2011; Mallinson et al., 2010, 2011; Mallinson et al., 2018; Zaremba et al.,
2016) each referenced to the present‐day mean sea level vertical datum. In particular, Culver et al. (2007)
provide the overall interpretation of sediment core data and describe the segmentation of the Outer Banks
barrier islands system twice during the late Holocene, with subsequent work (Grand Pre et al., 2011) that
describes the detailed geological observations, interpretations, and implications. Detailed descriptions of
the present‐day hydrodynamic observations are presented in Clunies et al. (2017) and the geological
observations are presented in detail in Zaremba et al. (2016). Paleo‐inlet locations and size are based on
data from Culver et al. (2007), Grand Pre et al. (2011), and Mallinson et al. (2010, 2011). The approach
adopted in the present study is to use a numerical model with high temporal resolution to elucidate the
hydrodynamics over a short time scale that results from the major differences in bathymetry that are
defined by geologic interpretation over a long time scale.
A12‐month time period (corresponding to 1 October 2007 to 30 September 2008) is used to examine the
winds, ocean water levels, and inflows from the four major rivers that drive the hydrodynamic conditions
in the APES. Hourly observations of wind components (u, v) observed at Diamond Shoals (DS; National
Data Buoy Center Buoy 41025) are shown in Figure 2a. Hourly water levels in the ocean were obtained from
National Data Buoy Center stations DUKN7 (FO) at Duck and BFTN7 (BF) at Beaufort (Figure 2b); locations
shown in Figure 1a. River discharges gauged at a temporal resolution of 15 min were obtained in the Neuse
River Estuary (NRE) at U.S. Geological Survey (USGS Station 02091814) near Fort Barnwell and in the Tar‐
Pamlico River Estuary (TPRE) near Washington (USGS Station 0204472) (shown in Figure 2c). Data for the
flow rates of the Roanoke and Chowan rivers are not shown for clarity; however, they are smaller than the
flows entering the NRE and TPRE and were input to the numerical model. Water levels and current
Figure 1. Bathymetry of the APES showing: (a) location of hydrodynamic observations (FO, BF, and P2), selected sediment cores (C1 and C2), wind observations
(DS), the Tar‐Pamlico River Estuary (TPRE), and the Neuse River Estuary (NRE); (b) constrained seismic grid (black lines) and sediment core sites (circles) in
Pamlico Sound.
10.1029/2018JC014732
Journal of Geophysical Research: Oceans
MULLIGAN ET AL. 9237

velocities were observed in Pamlico Sound using a Nortek Acoustic Doppler Velocimeter at P2 (5.0‐m mean
water depth), which burst sampled for 4 min out of every hour at a rate of 8 Hz and was deployed at selected
times in 2007–2008.
Salinity observations were obtained from three sources, discussed in more detail in sections 3.3 and 4.3. The
data sources include the University of North Carolina modeling and monitoring (MODMON) project with
observations throughout the water column at locations throughout the NRE; the North Carolina
Department of Environment and Natural Resources Pamlico River Water Quality Monitoring project
(PAMRIV) with measurements of the surface and near‐bottom salinity in the TPRE; and the University of
North Carolina Ferry Monitoring (FerryMon) project with surface water quality data collected on North
Carolina Department of Transportation ferries crossing Pamlico Sound.
3. Numerical Model
Delft3D is a three‐dimensional (3‐D) numerical modeling system consisting of several integrated models that
simulate hydrodynamics, transport of water‐borne constituents (e.g., salinity and heat), surface waves by
coupling with the SWAN (Simulating WAves Nearshore; Booij et al., 1999) spectral wave model, sediment
transport, and morphological changes. Delft3D has been used to simulate coastal hydrodynamics with con-
siderable success when compared to observations in different shallow coastal environments such as embay-
ments (Mulligan, Hay, et al., 2008), river deltas (Hu et al., 2009), and river mouths (Elias et al., 2012). It has
also been used to simulate storm surge in the APES (Mulligan et al., 2015), waves on estuarine shorelines in
Pamlico Sound (Eulie et al., 2016), and estuarine circulation in the NRE (Brown et al., 2014). Delft3D can be
run in 3‐D mode where topography‐following σlayers define the vertical coordinate. In this study, a 3‐D
model with vertical layers is used which allows for prediction of stratification and vertical mixing that are
important for simulating salinity. A detailed description of the system of equations and numerical imple-
mentation is provided by Lesser et al. (2004).
Figure 2. Observations over a 12‐month period from 1 October 2007 to 30 September 2008 that are input to the model: (a) wind components at DS (Diamond Shoals
Buoy); (b) ocean water levels at FO (FRF Ocean side) and BF (Beaufort); and (c) selected river discharges at the head of the NRE and TPRE. Site locations are shown
in Figure 1.
10.1029/2018JC014732
Journal of Geophysical Research: Oceans
MULLIGAN ET AL. 9238

3.1. Model Setup
The model domain is a curvilinear grid, with an open boundary along the continental shelf that has higher
resolution in the narrow subestuaries of the APES and lower resolution in the Atlantic Ocean. The resolu-
tion of the flow model grid ranges from 250–300 m, with the wave model grid having a resolution of
500–600 m. Both model domains covered an area of 220 km in the east‐west direction and 295 km in the
north‐south direction. The present‐day bathymetric grid was built using a high‐resolution digital elevation
model of Pamlico Sound, obtained from National Oceanic and Atmospheric Administration and the
USGS (Cross et al., 2005), at a resolution of 30 m. The vertical grid has eight vertical σlayers, and each layer
has a thickness equal to 12.5% of the local water depth. Since the APES has a maximum depth of 8 m and
typical depths of 1–5 m, this means the vertical resolution ranges from approximately 0.1–1.0 m in agree-
ment with other modeling studies in estuaries and inlets (e.g., Purkiani et al., 2015). Model simulations were
run for a 12‐month period from 1 October 2007 to 30 September 2008, with computations performed at a
time step of 30 s. Model spin‐up is an important consideration when initializing a numerical model. In this
case, the water levels and current speeds were near 0 at the start of the run and very good model agreement
with observations over the simulation period (Clunies et al., 2017). This indicates that the model spin‐up
time for water level elevations is very short (i.e., <2 days) for Pamlico Sound. The initial conditions for sali-
nity are very important since the system has a horizontal salinity gradient from fresh water to ocean water. A
realistic spatial distribution is used to initialize the model salinity field, discussed further in section 3.3. All
model runs use the same hydrodynamic forcing from winds, tides, and river inflows. Global tidal modeling
at 1,000‐yr intervals over the past 10,000 yr has shown that changes in the spatial patterns and amplitudes of
tidal constituents are small in the middle to late Holocene (Hill et al., 2011). It is therefore reasonable to
assume that tidal range at the ocean boundary has remained very similar over the last 4,000 yr, and any dif-
ferences in the ocean are significantly smaller compared to changes to the coastal/estuarine system. In this
approach, the hydrodynamic responses to different morphology of the barrier islands and estuary
are isolated.
The model runs were forced using a spatially uniform wind field, water levels at the open boundary, and
freshwater river discharges for the Neuse, Tar‐Pamlico, Chowan, and Roanoke Rivers (Figure 2c). The wind
field, described by Clunies et al. (2017), was developed from observations made at DS (Figure 2a) with hourly
temporal resolution. The open boundary condition was developed from water level observations at BF and
FO (Figures 1 and 2b) by linearly interpolating along the open boundary to develop a spatially varying
boundary condition. The simulations were initialized with a water level of 0 m (mean sea level). Model para-
meters and sensitivity tests are described in detail by Clunies et al. (2017) and were validated using available
observations of water levels, waves, and currents for the present‐day scenario. The bottom drag coefficient
(c
D
) was set to the default value of 0.0023, and a k‐ɛturbulence closure scheme was used with background
horizontal and vertical eddy viscosity coefficients of 1.0 and 1.0×10
−5
m
2
s
−1
, respectively. In SWAN, fre-
quency was defined using 49 logarithmically spaced bins from 0.05–3.00 Hz and directions were defined
using 36 bins with 10° spacing. Depth‐induced breaking in shallow water was defined using the bore‐based
model of Battjes and Janssen (1978), with the rate of dissipation coefficient (α) set to 1.0 and the breaker
parameter (γ) set to 0.73. Bottom friction was parameterized using the semiempirical formulae from the
JONSWAP experiments (Hasselmann et al., 1973), using a bottom friction coefficient (c
b
) equal to 0.067
m
2
s
−3
, appropriate for locally generated wind sea. Whitecapping was modeled based on the formulation
described by van der Westhuysen et al. (2007) and Mulligan, Bowen, et al. (2008).
3.2. Paleobathymetric Grids
Zaremba et al. (2016) describe the geological observations corresponding to the late Holocene stratigraphic
record and paleo‐environmental evolution of Pamlico Sound, with a focus on time periods of increased mar-
ine influence. These periods are interpreted to be the result of extensive barrier island segmentation synchro-
nous with periods of rapid climate change, indicating the highly dynamic character of the coastal system in
response to sea level changes and interactions with paleotopography. In the present numerical study, the
hydrodynamic and wave models are applied to four different geomorphic scenarios characterized by changes
to the bathymetry and inlet configurations and corresponding to distinct time intervals (present day, 500,
1,000, and 4,000 cal yr BP) during the evolution of Pamlico Sound as described by Zaremba et al. (2016).
The paleobathymetric grids were created using age‐depth relationships developed from cores in Pamlico
10.1029/2018JC014732
Journal of Geophysical Research: Oceans
MULLIGAN ET AL. 9239

Sound in combination with the time‐constrained seismic grid data with locations shown in Figure 1b. The
paleosurfaces implemented in the numerical model are therefore derived from the present‐day surface,
after modification based on observations from the sediment cores and geophysical surveys. Relative sea
level curves developed from geological sea level reconstructions (Kopp et al., 2015) were used as a
reference for water depth at the time of formation of significant stratigraphic horizons. SLR corrections
for the 500, 1,000, and 4,000 cal yr BP bathymetric grids were −0.61, −1.35, and −3.06 m, respectively in
comparison to the present‐day sea level. In the present, Pamlico Sound has three inlets shown in Figure 3
Figure 3. Bathymetry of Pamlico Sound for four different geomorphic scenarios: (a) present day; (b) 500 cal yr BP; (c) 1,000 cal yr BP; and (d) 4,000 cal yr BP
(adapted from the observations described by Zaremba et al., 2016). The 3‐and 5‐m depth contours are indicated by black lines, and the inlet locations are indi-
cated by red boxes. The vertical datum is present‐day mean sea level (MSL) for all scenarios.
10.1029/2018JC014732
Journal of Geophysical Research: Oceans
MULLIGAN ET AL. 9240

a: Ocracoke Inlet, Hatteras Inlet, and Oregon Inlet. In addition to changes made to the Pamlico Sound basin
morphology, the grids were modified to account for paleo‐inlets occurring during the different time intervals
based on the work of Culver et al. (2007), Grand Pre et al. (2011), and Mallinson et al. (2010, 2011)
(Figures 3b–3d). Paleo‐inlet dimensions (i.e., the active inlet throat channels) are difficult to estimate
based on geological data, because of the migration tendencies of the channel within the overall fill
sequence, as well as difficulties in imaging with ground penetrating radar greater than 4 m below the
subsurface because of the effects of salt water. Thus, for model purposes inlet dimensions utilized a
modern analogue (Drum Inlet) with an inlet throat channel width of 0.5 km and maximum depth of 6 m.
This is consistent with geological constraints provided by ground penetrating radar data (Mallinson et al.,
2010). Based on the inlet‐throat area to tidal prism relationship (Hughes, 2002; Jarrett, 1976), each inlet
would accommodate a tidal prism of 4.6 × 10
7
m
3
, which is approximately half of the average (1988–1989)
tidal prism at Oregon Inlet (1.1 × 10
8
m
3
; Nichols & Pietrafesa, 1997). Thus, model results likely represent
minimum expected changes within the Sound, given the relatively small size of modeled inlets. The
bathymetry in each geomorphic scenario is assumed to be in morphodynamic equilibrium since it is
derived from the geological record.
3.3. Initial Salinity Conditions
Advection‐diffusion of salinity is simulated using the 3‐D transport equation (Lesser et al., 2004) in horizon-
tal (x,y) and vertical topography‐following (σ) coordinates to determine the salt concentration in the model
domain. The horizontal and vertical diffusion coefficients have values of D
H
= 0.1 m
2
s
−1
and D
V
= 1.0×10
−5
m
2
s
−1
. The predicted salinity is compared to available data in September from the MODMON, PAMRIV,
and FerryMon monitoring projects at the stations shown in Figure 4a.
Simulating salinity in the estuaries requires the initial salinity conditions to be specified throughout the
model domain. Initial salinity values in the APES were developed for the model using typical fall salinity
values described by Wells and Kim (1989) that compare well with the model results of Jia and Li (2012).
These are shown in Figure 4b, with a strong horizontal gradient from the river inflows (0 psu) to central
Pamlico Sound (~20 psu). Ocean salinity outside the Outer Banks and along the offshore boundary was
set to a value of 34 psu to represent the North Carolina shelf water, and the initial salinity conditions
Figure 4. Estuarine salinity: (a) observation sites from the FerryMon, MODMON, and PAMRIV monitoring projects and (b) model domain and initial conditions at
the start of each model run.
10.1029/2018JC014732
Journal of Geophysical Research: Oceans
MULLIGAN ET AL. 9241

were depth‐uniform throughout the model domain. For computational efficiency, eight vertical σlayers are
used in this shallow system (average depth is 5 m), and therefore, each layer has an average thickness of 0.6
m. After initializing the model with the salinity distribution, it was determined that the spin‐up time is
approximately 6 months indicated by the adjustment to steady‐state salinity over the 12‐month
simulations. There is a marked difference in model spin‐up time scales for water levels (2 days) and for
salinity (6 months). This is because the tidal influence on water level elevations in Pamlico Sound is very
small compared to the larger wind‐driven water level fluctuations. However, salinity, as an indicator of
water mass transport, requires a much longer time to achieve a steady‐state balance between the
freshwater discharge from rivers that enter on the west side and tidal pumping of marine water through
inlets on the east side of the Sound.
4. Results
4.1. Hydrodynamics
Model validation for the present‐day scenario is explained in detail in Clunies et al. (2017), where water
level statistics including the correlation coefficients (R= 0.80–0.93) and root‐mean‐square errors (RMSE
= 0.06–0.10 m) are provided over a 35‐day simulation at six sites. For the significantly longer model
Figure 5. Model validation over the 12‐month simulation period in Pamlico Sound (Site P2): (a) water level elevation (RMSE = 0.08 m), with selected period shown
by the gray shaded box; (b) current magnitude (RMSE = 0.04 m/s); and (c and d) detail of water levels and currents during selected period in September 2008.
10.1029/2018JC014732
Journal of Geophysical Research: Oceans
MULLIGAN ET AL. 9242

simulation in the present study, validation for water levels and currents in Pamlico Sound is presented in
Figure 5 using observations at Site P2 over a 1‐yr period. These results have low errors for water levels
(RMSE = 0.08 m) and for currents (RMSE = 0.04 m/s), which indicates high confidence in the
hydrodynamic numerical results. The water levels and currents in Pamlico Sound have a minimal tidal
influence and the variability is almost exclusively wind driven, in agreement with other studies
(Clunies et al., 2017; Luettich et al., 2002).
Figure 6. Simulated water level elevations above mean sea level at high tide (corresponding to the conditions on 25 September at 1230, for the four different geo-
morphic scenarios: (a) present day; (b) 500 cal yr BP; (c) 1,000 cal yr BP; and (d) 4,000 cal yr BP.
10.1029/2018JC014732
Journal of Geophysical Research: Oceans
MULLIGAN ET AL. 9243

The response of Pamlico Sound to wind and ocean water level forcing is investigated for distinct scenarios,
and the influence of the different morphological conditions is evaluated through the comparison between
the model results for each time slice. Simulated water levels are shown in Figure 6 at a time with high tidal
elevations in the ocean, and the differences between the four time slices indicate the influence of the differ-
ent inlet configurations on water levels in the Sound. Similar to the water level results, current velocity pre-
dictions indicate a different tidal signal in each case. Maximum simulated current velocities in the 4,000 cal
Figure 7. Simulated velocities during maximum flood tidal current (corresponding to the conditions on 2 August at 0030) for the four different geomorphic scenar-
ios: (a) present day; (b) 500 cal yr BP; (c) 1,000 cal yr BP; and (d) 4,000 cal yr BP.
10.1029/2018JC014732
Journal of Geophysical Research: Oceans
MULLIGAN ET AL. 9244

yr BP model are strongest near the inlets (e.g., near Site C2), reaching approximately 0.8 m s
−1
, 3 times
higher than current velocities in the present‐day model at C2 and indicating the importance of inlets on
circulation in the back‐barrier estuary.
The flows during maximum flood‐tidal currents with calm winds (U
w
<5ms
−1
) are shown in Figure 7. The
greater number of inlets (500 cal yr BP, Figure 3b) and wider inlets (1,000 cal yr BP, Figure 3c) result in
higher current velocities in Pamlico Sound, with a noticeable increase in velocity in central Pamlico
Sound (1,000 and 4,000 cal yr BP) and the mouths of the TPRE and NRE (4,000 cal yr BP, Figure 3d). The
flow during maximum ebb‐tidal currents (Figure 8) through the inlets indicates higher current velocities
Figure 8. Simulated velocities during maximum ebb tidal current (corresponding to the conditions on 22 September at 1230) for four different geomorphic scenar-
ios: (a) present day; (b) 500 cal yr BP; (c) 1,000 cal yr BP; and (d) 4,000 cal yr BP.
10.1029/2018JC014732
Journal of Geophysical Research: Oceans
MULLIGAN ET AL. 9245

in Pamlico Sound than in the present‐day case, even at considerable distances from the inlets. This is caused
by higher flow rates through the system due to the different inlet configurations, since the inlets are either
wider (e.g., 21 km wide at 1,000 cal yr BP) or there is a greater number of smaller inlets (e.g., nine inlets at
500 cal yr BP) and is not restricted to the three inlets in the present‐day scenario (Figure 3a).
Currents in central Pamlico Sound are also influenced by the mean water depth change, particularly over the
shallow north‐south oriented sand ridge called Bluff Shoal. Bluff Shoal is a relic interfluvial divide between
the Neuse River and Pamlico/Tar River systems, dating prior to 7,500 cal yr BP (Zaremba et al., 2016) and is a
good example of the role of antecedent topography on hydrodynamics (Mallinson et al., 2010). SLR of
approximately 3 m over the last 4,000 yr (Horton et al., 2009; Kemp et al., 2009, 2017) has increased the depth
over Bluff Shoal and decreased the current magnitude over this feature. In the 4,000 and 1,000 cal yr BP
cases, with lower sea level, a deviation in flow direction occurs at Bluff Shoal during the ebb tide. Bluff
Shoal impedes the flow of water driven by winds and tides, and water from Pamlico Sound is able to exit
the back‐barrier environment through the closest paleo‐inlets near the present‐day site of Ocracoke Island.
To compare the hydrodynamic results for each scenario, water levels and current velocities for a short time
near the end of the simulations are presented in Figure 9 at three locations (P2, C1, and C2; Figure 1a) in
Pamlico Sound for morphological conditions corresponding to each geomorphic scenario. During each time
period a strong northerly wind event drives a storm surge into southwestern Pamlico Sound and the wind‐
driven water levels are significantly greater than the tides in the estuary. The results of the present‐day run
are in good agreement with the water level and current observations at P2, indicating microtidal conditions
Figure 9. Comparison of selected time series of model results (water level elevation, depth‐averaged current magnitude, and salinity) for the four simulations at
Sites P2, C1, and C2 corresponding to the conditions over a 10‐day period in September 2008.
10.1029/2018JC014732
Journal of Geophysical Research: Oceans
MULLIGAN ET AL. 9246

with a tidal range of <0.10 m, storm surge of 0.60 m, and maximum current speeds of 0.10 m s
−1
. The inlets
along the Outer Banks have a significant effect on the simulated water levels and current velocities in
Pamlico Sound at C1 and C2 (Figure 9). The water levels indicate higher tidal energy in Pamlico Sound,
in comparison to the present‐day model since tides are able to propagate farther into the back‐barrier envir-
onment. The tidal signal is strongest at the observation sites (C1 and C2) in the 4,000 cal yr BP scenario due
to the proximity of these sites to inlets at this time. The average tidal range, the residual after removing the
larger wind‐driven surge, in Pamlico Sound (at P2) is approximately 0.30, 0.20, 0.15, and 0.10 m for the 4,000,
1,000, 500, and present‐day geomorphic scenarios, respectively.
4.2. Surface Waves
Validation of the wave model for the present‐day scenario (Clunies et al., 2017) in Albemarle Sound indi-
cated the wave climate is fetch or depth limited in the estuarine system depending on the wind direction.
Significant wave height (H
s
) for the present day and 1,000 cal yr BP geomorphic scenarios, for strong winds
blowing from the south (Figures 10a and 10c) and the north (Figures 10b and 10d), indicate differences in
wave energy of 10–20%. For winds from the south, the 1,000 cal yr BP simulation had the largest significant
wave heights, with waves up to H
s
= 1.5 m in Pamlico Sound near the mouths of the NRE and TPRE, with
some wave energy from the ocean that is transmitted into the sound through two inlets. Similarly, the trans-
mission of wave energy from the ocean into Pamlico Sound also occurs in the 1,000 cal yr BP run through an
inlet west of Cape Hatteras, increasing the wave height in eastern Pamlico Sound. Wave energy is also trans-
mitted through the wide inlet in the 1,000 cal yr BP geomorphic scenario into central Pamlico Sound but is
reduced by depth‐limited breaking in the shallow sound. Wave energy from the ocean through these two
inlets during south winds results in H
s
= 1.4 m in the western region of the northern basin of Pamlico
Sound. The smaller inlets in the present‐day case and the relatively shallow inlets in the 500 cal yr BP geo-
morphic scenario limit wave energy in Pamlico Sound to approximately H
s
< 1.2 m for these wind condi-
tions. For winds from the north, the present‐day geomorphic scenario has the largest H
s
, reaching 1.5 m
in a significant portion of Pamlico Sound due to the greater water depth. The smaller number of narrower
inlets in the present‐day case results in negligible transmission of wave energy from Pamlico Sound to the
ocean. The wider and deeper inlets in the paleobathymetric model runs allow a higher amount of wave
energy transmission from the sound to the ocean.
4.3. Salinity
To assess the changes in salinity distributions in Pamlico Sound and exchange between the ocean and the
back‐barrier environment, the model was used to simulate the salt concentration for each geomorphic sce-
nario. The simulations do not intend to predict the exact salinity distributions but highlight important differ-
ences in circulation for the different inlet configurations. However, the salinity predictions are first validated
using observations in the present day in three regions of the estuary: the NRE, the TPRE, and Pamlico Sound
along a ferry route (Figure 4a). Observations and model results are shown in the NRE and TPRE for one day
(29 September 2008) at the end of the simulation. These results indicate that the salinity field has some rea-
listic aspects that match the field measurements. First, the lowest simulated salinity values occur at the
heads of the subestuaries near the freshwater sources from the Neuse and Tar Rivers with the estuaries
becoming higher in salinity eastward toward Pamlico Sound. Second, stratification is much stronger in
the NRE than in the TPRE due to higher discharge in the NRE, indicated by the observed and predicted sur-
face (Figure 11a) and near bed level (Figure 11b) salinity values. The vertical structure of salinity in the NRE
is apparent in the model (Figure 11c), although the halocline is slightly deeper and salinity values are lower
near the bottom in the model. Overall, the error of the salinity at the NRE surface has a spatially averaged
RMSE = 2.4 psu and at the bottom RMSE = 9.6 psu. This error in near‐bottom salinity after 1 yr of simulation
could possibly be reduced by increasing the vertical resolution and improving the simulation of baroclinic
processes; however, this is computationally intensive and the surface layer has lower error in both the
NRE and the TPRE. The TPRE is vertically well mixed (Figure 11d) and the surface salinity error in the
TPRE is RMSE = 3.4 psu, and at the bottom it is lower with RMSE = 1.5 psu. Observations and model results
indicate that the water column is well mixed in Pamlico Sound indicated by the near identical salinity at the
NRE and TPRE estuary mouths in Figure 11.
The variation in the simulated surface salinity distribution and the FerryMon observations on 12 different
days in September 2008 are shown in Figure 12. The model results for surface salinity are not in perfect
10.1029/2018JC014732
Journal of Geophysical Research: Oceans
MULLIGAN ET AL. 9247

agreement with FerryMon measurements across the southeastern part of Pamlico Sound, likely due to the
highly dynamic exchange of water masses in this region. However, the model results capture the
exchange of high salinity ocean water with the brackish estuarine water through Ocracoke Inlet and
variability over time. For example, export of lower salinity water (26–30 psu) through the inlet to the
ocean occurs in the model on several days (e.g., Figure 12a) and plumes of higher salinity water (31–35
psu) occur inside the sound in the model and observations (e.g., Figure 12d). The model predicts the
Figure 10. Simulated significant wave height in Pamlico Sound for (a) 17‐ms
−1
winds from the south (6 September at 1230) and (b) 16‐ms
−1
winds from the north
(24 September at 0030) for basin morphology corresponding to the present day; and the same winds from (c) the south and (d) the north for basin morphology
corresponding to the 1,000 cal yr BP scenario.
10.1029/2018JC014732
Journal of Geophysical Research: Oceans
MULLIGAN ET AL. 9248

Figure 11. Salinity observations and model results in the NRE and TPRE on 29 September 2008: (a) near‐surface model results and observations (colored circles);
(b) near‐bottom model results and observations; (c) transect along the TPRE; and (d) transect along the NRE.
10.1029/2018JC014732
Journal of Geophysical Research: Oceans
MULLIGAN ET AL. 9249

variability in salinity distribution across the region with differences from observations that have RMSE
ranging from 1.4–3.6 psu for a given day and an average RMSE = 2.1 psu for all data shown. This suggests
that the model is a useful tool to investigate salinity differences and identify important trends between the
geomorphic scenarios, though modeled salinity values may not be precise. The model is applied to each
geomorphic scenario to determine realizations of the salinity distribution over a 12‐month period with the
same realistic hydrodynamic forcing conditions.
Time‐averaged (6‐month) spatial distributions of near‐bed salinity in the estuary are shown in Figure 13.
These model results indicate that the lowest salinity (20–25 psu) in central Pamlico Sound occurs in the
present‐day geomorphic scenario (Figure 13a), due to the fewer and smaller inlets. Results for the 500 and
1,000 cal yr BP geomorphic scenarios indicate higher salinity (25–34 psu) in central Pamlico Sound, the
strong influence of the very shallow Bluff Shoal in dividing Pamlico Sound into two basins, and lower sali-
nity in western regions of Pamlico Sound with marine salinity water in the eastern region near the paleo‐
inlets open along the Outer Banks (Figures 13b and 13c). Results for the 4,000 cal yr BP time slice also
Figure 12. Simulated estuary surface salinity compared to FerryMon salinity observations (colored circles) on 12 days in September 2008. The RMSE ranges from
1.4–3.6 psu for each day, and the average RMSE of all data shown is 2.1 psu.
10.1029/2018JC014732
Journal of Geophysical Research: Oceans
MULLIGAN ET AL. 9250

indicate high salinity in Pamlico Sound compared to the present‐day case. The southern inlet at this time
also allows greater transport of estuarine water to the ocean. Salinity ranges developed from observed
foraminiferal assemblages in sediment cores are superimposed on the salinity distributions in Figure 13.
Variability in salinity is also shown in the time series at selected sites in Figure 9. Although the salinity
values interpreted from core data represent long‐term environmental averages and the model results
indicate changes over a 6‐month average period, this approach demonstrates the effect of different inlets
on water mass properties in the estuary and exchange with the ocean.
Figure 13. Six‐month averaged (April–September) maps of simulated near‐bed salinity for four geomorphic scenarios: (a) present day; (b) 500 cal yr BP; (c) 1,000 cal
yr BP; and (d) 4,000 cal yr BP (see color bar). Long‐term approximate mean salinity values determined from foraminiferal assemblages at sediment cores in Pamlico
Sound are shown by the colored circles (see legend).
10.1029/2018JC014732
Journal of Geophysical Research: Oceans
MULLIGAN ET AL. 9251

5. Discussion
5.1. Hydrodynamic Responses to Long‐Term Changes in Basin Morphology
Model simulations for the three geomorphic scenarios indicate that more inlets or wider inlets result in
higher currents even at large distances from the inlets. Wider inlets allow more tidal energy to enter the
basin and therefore increase velocities in the basin (farther from the inlets); however, wider inlets would
have lower currents in the inlet channels for the same tidal conditions. Although the 1,000 cal yr BP geo-
morphic scenario has the largest inlet (21 km), the highest water levels and currents in central Pamlico
Sound occur in the 4,000 cal yr BP simulation. SLR and inlet dynamics (breaching and closure) have also
changed the wave climate in Pamlico Sound over the last 4,000 yr. In the present day, waves from the ocean
are unable to propagate into the back‐barrier system due to the relatively small size of the inlets (Figure 3a).
The results in Figure 10 indicate that for different wind directions, barrier segmentation (in combination
with geomorphic changes due to long‐term sediment transport and deposition in the basin) has a strong
influence on the estuarine wave climate. The results suggest that changes in water depth due to SLR have
an impact on the wave climate for winds from the north, with the largest significant wave heights in the east-
ern basin of Pamlico Sound for the present‐day setting (Figure 10a). For winds blowing from the south, the
barrier island configuration and basin morphology become stronger controls on the wave climate. When
compared to the present day, the 4,000 and 1,000 cal yr BP cases have smaller basins in central Pamlico
Sound and a shallower depth over Bluff Shoal, which cause depth‐limited breaking. The wave height results
also suggest that the inlets in the 1,000 cal yr BP geomorphic scenario, in comparison to smaller inlets in the
present‐day case, have a greater effect on the wave climate. These results provide an example of the syner-
gistic effects of barrier segmentation and long‐term bathymetric changes on significant wave heights in a
back‐barrier estuary such as Pamlico Sound. The wider Ocracoke Inlet in the 1,000 cal yr BP case allows a
higher amount of wave energy to enter the back‐barrier estuary from the ocean, resulting in higher signifi-
cant wave heights in Pamlico Sound (Figure 10d) than occur in the present‐day (Figure 10b) barrier island
configuration. Greater wave heights would likely enhance estuarine shoreline erosion rates, increasing sedi-
ment flux to the basin, and also more effectively winnowing those sediments and preferentially exporting
mud from the system to the shelf via the inlets. This agrees with the observed increase in the sand/mud ratio
of sediments corresponding approximately to the 1,000 cal yr BP time interval (Zaremba et al., 2016).
5.2. Controls on Salinity Distributions
The model results represent relatively short (1‐yr) duration realizations of the hydrodynamic response to the
different bathymetry in each geomorphic scenario. The simulations include freshwater inflows from the
major rivers using measured data (Figure 2) but do not include fluctuations in salinity resulting from
longer‐term changes in rainfall or runoff patterns. This is apparent when comparing the foraminifera‐
derived salinity to the salinity model for the present day (Figure 13a). The foraminiferal data in this case
are from near the top of the various cores and reflect conditions between approximately 70 and 400 yr ago
(based on average accumulation rates of 1 mm/yr). Geochemical and microfossil data indicate that marine
influence within Pamlico Sound has increased over the last century (Abbene, 2004; Abbene et al., 2006);
thus, the foraminiferal assemblages reflect lower salinity conditions of the last several centuries, whereas
the model reflects modern conditions. An additional complicating factor is that, because of salinity toler-
ances, the foraminifera‐based salinity estimates reflect conditions within a range of approximately 5 to 10
psu. Furthermore, the values can be considered as minima as the presence and relative abundance of the
species involved are limited more by lower (estuarine) rather than higher (i.e., estuarine, but closer to nor-
mal marine, or normal marine) salinity values. This demonstrates that the models are not directly compar-
able with the centennial‐to millennial‐scale changes inferred from the depositional record preserved in the
sediment cores but are still useful for interpreting the response of the estuary to the different geomorpholo-
gical conditions. With these caveats in mind, the models generally support the higher salinity trends sug-
gested by distributions of foraminifera during the 4,000 (in particular) and 1,000 cal yr BP time intervals
and lower salinities of the present day (Figure 13; Foley, 2007; Metger, 2009; Grand Pre et al., 2011). An
exception to this is the simulated high salinity conditions within Pamlico Sound during the 500 cal yr BP sce-
nario, which are not supported by microfossil data (Figure 13b), in contrast to the salinity conditions during
the 4,000 and 1,000 cal yr BP periods. This suggests that tidal exchange was not as high at 500 cal yr BP as
during earlier and more segmented scenarios; thus, inlets were likely smaller and more ephemeral and
10.1029/2018JC014732
Journal of Geophysical Research: Oceans
MULLIGAN ET AL. 9252

the number of inlets open at one time in the geomorphic reconstruction may be an overestimate. Salinities
were higher at 1,000 cal yr BP than present by 5–10 psu and at 4,000 cal yr BP salinities were slightly lower
than at 1,000 cal yr BP, suggesting differences in connections with the ocean during these periods.
The simulated surface salinity distributions suggest that water properties in Pamlico Sound are controlled by
mixing due to winds and waves, the degree of connectivity (i.e., flow‐through area) between the higher sali-
nity Pamlico Sound and lower salinity Albemarle Sound, and the number and size of inlets and the strength
of the currents that influence flows in Pamlico Sound. Albemarle Sound is connected to Pamlico Sound via
the shallow Roanoke and Croatan Sounds, with Albemarle Sound having considerably lower salinity water
than Pamlico Sound (Wells & Kim, 1989). Apart from the net seaward flow that occurs as a result of the dis-
charge from river inputs, wind is the governing force that controls the salinity distribution in Pamlico Sound.
Changes in monthly, seasonal, decadal, or longer time scale winds as a result of climate change could dras-
tically affect the transport of water between Albemarle and Pamlico Sound and influence the
salinity distribution.
Connectivity between Albemarle and Pamlico Sounds was likely nonexistent during lower sea level condi-
tions prior to ~500 cal yr BP due to the occurrence of an interfluvial divide and an extensive marsh system,
in the location of modern‐day Croatan Sound (Figure 1a). Connectivity increased greatly during the early
nineteenth century when inlet closure along the northern Outer Banks forced water south through
Croatan Sound, removing the marshes and rapidly expanding this connection (Riggs et al., 2000; Riggs &
Ames, 2003). The lack of this connectivity between the sounds at lower sea levels relative to the present
day is illustrated in Figures 13b–13d, where salinity along the northern shoreline of Pamlico Sound was
much greater during the three paleoscenarios compared to the present day.
In all geomorphic scenarios, a net seaward transport of water from Pamlico Sound into the Atlantic Ocean
through the inlets occurs as a result of flows from the major rivers. Changes in the size, number, and location
of inlets along the Outer Banks create alternative flow paths for water through the estuary, leading to higher
flushing rates and shorter residence times relative to the present day. Plumes of lower salinity water that are
typically confined to the northern region of Pamlico Sound in the present‐day case are able to exit the back‐
barrier environment via paleo‐inlets located along Hatteras Island (e.g., in the 500 and 1,000 cal yr BP geo-
morphic scenarios). The paleobathymetric model results indicate that morphological changes, including
changes to the bathymetry, shoreline position, opening of new inlets, changes to inlet size, or changes to
inlet location have a significant and complex influence on the hydrodynamic energy and salinity distribution
in Pamlico Sound. The changes are complex and include morphodynamic feedbacks via sediment transport
such that they are not easily forecast without a numerical modeling approach. Sediment transport modeling
is an important next step; however, it would require a different method consisting of significantly longer
model runs or “morphologic upscaling”(e.g., Ranasinghe et al., 2011) over long time scales.
5.3. Insight From New Results and Comparison With Other Back‐Barrier Systems
Studies of other large back‐barrier systems indicate the sensitivity of these regions to SLR. For example,
Beets and van der Spek (2000) investigated the evolution of the back‐back barrier evolution of estuaries in
Belgium and the Netherlands during the sea level changes over the Holocene. Their study importantly found
differences in the balance between accommodation space and the sediment supply that led to stabilization of
the barrier islands and deposition in the Belgian basin, while higher tidal energy and sediment transport
capacity have kept the Netherlands large inlets and Wadden Sea open to present. Basin evolution in the
Wadden Sea has also been investigated numerically using Delft3D by Dissanayake et al. (2012) using an
idealized modeling approach over a 110‐yr future period to determine possible erosion/accretion rates.
This system has also been studied by Wang et al. (2018) for future SLR corresponding to various climate
change scenarios, to determine rates of sand sediment import into the tidal basins by tidal currents. In other
systems, Leorri et al. (2011) investigated multiple static sea level conditions in Delaware Bay, finding that
significant changes in the tidal range occur due to amplification or attenuation of the tides over different
bathymetry. Passeri et al. (2016) used a numerical model to simulate the tidal hydrodynamics of different
back‐barrier bays in the Gulf of Mexico for different SLR scenarios and emphasized the need to consider
the coevolution of morphology in conjunction with SLR for comprehensive evaluations of hydrodynamics.
Passeri et al. (2018) considered hurricane storm surge combined with SLR, quantifying the higher inunda-
tion and overwash that occurs when sea levels are higher in the Gulf of Mexico.
10.1029/2018JC014732
Journal of Geophysical Research: Oceans
MULLIGAN ET AL. 9253

Taken together with the new results of the present study, it is apparent that barrier systems are sensitive to
change (e.g., SLR or storms) and back‐barrier estuaries respond strongly and with great complexity, causing
regime shifts in hydrodynamics (water levels, currents, and waves) and water mass properties such as sali-
nity. The present method allows for an understanding of the mechanisms responsible for past changes that
are fundamentally grounded in reality and have several important implications for future changes. First, our
results indicate that relatively small perturbations to the morphology of the system such as opening a new
inlet or changing the depth can have major impacts on water levels that would drive flooding of the main-
land. This would have significant implications on the sediment distribution and dynamics within the sound
as circulation patterns and current strength drive sediment transport in the back‐barrier system. In the event
of barrier segmentation caused by new inlet formation in the future, reworking of the present‐day sediment
distribution within Pamlico Sound is expected. The results of this study improve understanding of the hydro-
dynamic response of the system to the changes in morphology that have already occurred as interpreted
from the geological record, and may also help inform studies of future change and management decisions.
Second, our results suggest that SLR could further increase the connectivity between Albemarle and
Pamlico Sounds in the future. This could result in a system‐wide shift in the salinity distribution with
impacts on habitat and the organisms that reside within the back‐barrier environment, for example, fish spe-
cies dependent on this region for spawning, larval, and juvenile stages of life. Third, with the possibility of
potentially accelerating SLR (Bernstein et al., 2007; Jevrejeeva et al., 2010, 2012; Rahmstorf et al., 2012),
in combination with projected increasing tropical cyclone intensity (Elsner et al., 2008; Emanuel, 2005),
the Outer Banks barrier islands and similar barrier‐island systems are at an increased risk for overwash, ero-
sion, and breaching.
6. Conclusions
Pamlico Sound is a large back‐barrier estuary that has evolved over the late Holocene in response to SLR,
similarly to other large systems such as the Wadden Sea in the Netherlands and many smaller back‐barrier
bays along the northern Gulf of Mexico coast and other parts of the world. The present study uses a coupled
numerical flow (Delft3D) and wave model (SWAN) with different bathymetric grids corresponding to four
distinct times during the geomorphic evolution of Pamlico Sound to assess the hydrodynamic, surface wave,
and salinity response to long‐term changes in geomorphology, barrier segmentation, and SLR. The presence
of inlets along the Outer Banks and changes to bathymetry in the model runs had a significant effect on the
water levels and current velocities in Pamlico Sound. Increasing the number and size of the inlets in the
paleobathymetric simulations resulted in increased tidal range in Pamlico Sound. Tidal signals were stron-
gest in the 4,000 cal yr BP scenario, with current magnitudes up to 3 times higher (0.78 m s
−1
) in comparison
to present‐day maximum currents of 0.25 m s
−1
. Higher current velocities in Pamlico Sound occurred due to
a greater number of inlets and due to wider inlets, and stronger currents were not limited to the areas adja-
cent to the inlets. Current velocities increased throughout the estuary in all three paleobathymetric simula-
tions during both flood and ebb tides suggesting a higher transport rate in the sound and exchange rate
between the sound and the ocean. Changes in inlet size and water depth due to sediment infilling and
SLR also impact the wave climate in Pamlico Sound, which would also influence erosion and transport of
estuarine sands and muds. Mean water level change due to SLR and sedimentation factors can therefore
influence the wave climate when winds are from the north as the waves are fetch and depth limited for
winds from this direction. Barrier segmentation and long‐term bathymetry changes more strongly affect
the wave climate when winds are from the south due to incident wave energy from the Atlantic Ocean that
is regulated by inlet size, depth, and abundance. Overall, the results indicate that with greater segmentation
comes greater tidal range and higher wave energy in the back‐barrier environment, which will accelerate
flooding and shoreline erosion.
The distribution of salinity throughout the estuarine system was computed for each geomorphic scenario.
The present‐day simulated salinity distribution was compared with observations from three monitoring pro-
jects and errors quantified that indicate high confidence in the model results for salinity. The comparisons
between each geomorphic scenario suggest that transport and mixing due to winds and waves, hydraulic
connectivity between Albemarle and Pamlico Sounds, and the number and size of inlets are three major fac-
tors influencing salinity in Pamlico Sound. The model results are not directly comparable with the 100‐to
1,000‐yr changes inferred from the microfossils (i.e., foraminiferal assemblages) preserved in the sediment
10.1029/2018JC014732
Journal of Geophysical Research: Oceans
MULLIGAN ET AL. 9254

cores since the models are short‐duration averages of the evolving hydrodynamics associated with each geo-
morphic scenario. Nevertheless, the modeling approach is useful for interpreting and comparing the bulk
response of the estuary to the different geomorphological conditions that occurred in the past. In the present,
Pamlico Sound grades from a partially mixed to a well‐mixed brackish estuary from east to west. The geolo-
gical observations and hydrodynamic modeling indicate that at certain times in the past, Pamlico Sound had
greater marine influence with greater current velocities, tidal ranges, and salinities. Given likely future sce-
narios of coastal evolution, for example, changes in water depth and inlets due to SLR and storms, Pamlico
Sound could return to a more marine‐influenced system. The introduction of greater tidal range, stronger
tidal currents, and increased wave energy will accelerate mainland flooding and shoreline erosion and have
profound impacts on estuarine shorelines and ecosystems, which will require a numerical modeling
approach for forecasting future changes.
References
Abbene, I. (2004). Reconstruction of the recent paleoenvironment of Pamlico Sound, North Carolina, using foraminifera, stable isotopes,
and radionuclide data. Master's Thesis, East Carolina University, North Carolina, pp. 1–158.
Abbene, I., Culver, S. J., Corbett, D. R., Buzas, M., & Tully, L. (2006). Distribution of foraminifera in Pamlico Sound, North Carolina, over
the past century. Journal of Foraminiferal Research,36, 135–151.
Battjes, J., & Janssen, J. (1978). Energy loss and set‐up due to breaking of random waves. Proceedings 16th International Conference Coastal
Engineering, ASCE, 569–587.
Beets, D. J., & van der Spek, A. J. (2000). The Holocene evolution of the barrier and the back‐barrier basins of Belgium and the
Netherlands as a function of late Weichselian morphology, relative sea‐level rise and sediment supply. Netherlands Journal of
Geosciences,79,3–16.
Bernstein, L., Bosch, P., Canziani, O., Chen, Z., Christ, R., Davidson, O., et al., (2007). Climate Change 2007: Synthesis Report. Valencia,
Spain, 26–73.
Bilskie, M. V., Hagen, S. C., Medeiros, S. C., & Passeri, D. L. (2014). Dynamics of sea level rise and coastal flooding. Geophysical Research
Letters,41, 927–934. https://doi.org/10.1002/2013GL058759
Booij, N., Ris, R. C., & Holthuijsen, L. H. (1999). A third‐generation wave model for coastal regions: 1. Model description and validation.
Journal of Geophysical Research,104(C4), 7649–7666.
Brown, M. M., Mulligan, R. P., & Miller, R. L. (2014). Modeling the transport of freshwater and dissolved organic carbon in the Neuse River
Estuary, NC, USA following Hurricane Irene (2011). Estuarine, Coastal and Shelf Science,139, 148–158.
Clunies, G., Mulligan, R. P., Mallinson, D., & Walsh, J. P. (2017). Modeling hydrodynamics of large lagoons: Insights from the Albemarle‐
Pamlico Estuarine System. Estuarine, Coastal and Shelf Science,189,90–103.
Cross, V., Bratton, J., Bergeron, E., Meunier, J., Crusius, J., & Koopmans, D. (2005). Continuous resistivity profiling data from the upper
Neuse River Estuary, North Carolina, 2004–2005. Open‐File Rep. 2005‐1306, USGS, Reston, VA.
Culver, S. J., Ames, D. V., Corbett, D. R., Mallinson, D. J., Riggs, S. R., Smith, C. G., & Vance, D. J. (2006). Foraminiferal and sedimentary
record of late Holocene barrier island evolution, Pea Island, North Carolina: The role of storm overwash, inlet processes, and anthro-
pogenic modification. Journal of Coastal Research,224, 836–846.
Culver, S. J., Grand Pre, C. A., Mallinson, D. J., Riggs, S. R., Corbett, D. R., Foley, J., et al. (2007). Late Holocene barrier island collapse:
Outer Banks, North Carolina, USA. The Sedimentary Record,5(4), 4–8.
Dissanayake, D. M. P. K., Ranasinghe, R., & Roelvink, J. A. (2012). The morphological response of large tidal inlet/ basin systems to relative
sea level rise. Climatic Change,113(2), 253–276.
Donnelly, J. P., Hawkes, A. D., Lane, P., MacDonald, D., Shuman, B. N., Toomey, M. R., et al. (2015). Climate forcing of unprecedented
intense‐hurricane activity in the last 2000 years. Earth's Future,3,49–65. https://doi.org/10.1002/2014EF000274
Donnelly, J. P., & Woodruff, J. D. (2007). Intense hurricane activity over the past 5,000 years controlled by El Nino and the West African
monsoon. Nature,447(7143), 465–468. https://doi.org/10.1038/nature05834
Elias, E. P., & van der Spek, A. J. (2006). Long‐term morphodynamic evolution of Texel Inlet and its ebb‐tidal delta (The Netherlands).
Marine Geology,225(1–4), 5–21.
Elias, E. P. L., Gelfenbaum, G., & van der Westhuysen, A. J. (2012). Validation of a coupled wave‐flow model in a high‐energy setting: The
mouth of the Columbia River. Journal of Geophysical Research,117, C09011. https://doi.org/10.1029/2012JC008105
Elias, E. P. L., & van der Spek, A. J. F. (2006). Long‐term morphodynamic evolution of Texel Inlet and its ebb‐tidal delta (The Netherlands).
Marine Geology,225(1‐4), 5–21.
Elsner, J. B., Kossin, J. P., & Jagger, T. H. (2008). The increasing intensity of the strongest tropical cyclones. Nature Letters,455. https://doi.
org/10.1038/nature07234
Emanuel, K. (2005). Increasing destructiveness of tropical cyclones over the past 30 years. Nature Letters,436. https://doi.org/10.1038/
nature03906
Eulie, D. O., Walsh, J. P., Corbett, D. R., & Mulligan, R. P. (2016). Temporal and spatial dynamics of estuarine shoreline change in the
Albemarle‐Pamlico estuarine system, North Carolina, USA. Estuaries and Coasts,40(3), 741–757. https://doi.org/10.1007/s12237‐016‐
0143‐8
Foley, J. (2007). Foraminiferal, sedimentological and geochemical indications of Holocene environmental change in Pamlico Sound.
Master's Thesis, East Carolina University, North Carolina, pp. 1–167.
González, J. L., & Törnqvist, T. E. (2009). A new Late Hol ocene sea‐level record from the Mississippi Delta: Evidence for a climate/sea level
connection? Quaternary Science Reviews,28,1737–1749.
Grand Pre, C., Culver, S. J., Mallinson, D. J., Farrell, K. M., Corbett, D. R., Horton, B., et al. (2011). Rapid Holocene coastal change
revealed by high‐resolution micropaleontological analysis, Pamlico Sound, North Carolina, USA. Quaternary Research,76(3),
319–334.
Harley, C. D. G., Randall Hughes, A., Hultgren, K. M., Miner, B. G., Sorte, C. J. B., Thornber , C. S., et al. (2006). The impacts of climate
change in coastal marine systems. Ecology Letters,9(2), 228–241. https://doi.org/10.1111/j.1461‐0248.2005.00871.x
10.1029/2018JC014732
Journal of Geophysical Research: Oceans
MULLIGAN ET AL.
Acknowledgments
Thanks to J. P. Walsh and D. Reide
Corbett at ECU for providing
hydrodynamic observations and
Benjamin Peierls at the University of
North Carolina at Chapel Hill and Jill
Paxson at the NC Department of
Environment and Natural Resources
for providing salinity data. The data
used in this study are archived in the
Department of Civil Engineering at
Queen's University and are available in
a data repository accessible online (doi:
10.5683/SP2/I6XNB8). The model
simulations were run using advanced
research computing resources provided
by Compute Canada. This study of
Coastal Hydrodynamics and Natural
Geologic Evolution (CHaNGE) was
funded by the U.S. National Science
Foundation under grant number OCE‐
1130843.
9255

Hasselmann, K., Barnett, T., Bouws, E., Carlson, H., Cartwright, D., Enke, K., et al., (1973). Measurements of wind‐wave growth and swell
decay during the Joint North Sea Wave Project (JONSWAP). Technical Report: Ergänzungsheft zur Deutschen Hydrographischen
Zeitschrift.
Hill, D. F., Griffiths, S. D., Peltier, W. R., Horton, B. P., & Törnqvist, T. E. (2011). High‐resolution numerical modeling of tides in the
western Atlantic, Gulf of Mexico, and Caribbean Sea during the Holoc ene. Journal of Geophysical Research,116, C10014. https://doi.org/
10.1029/2010JC006896
Horton, B. P., Peltier, W. R., Culver, S. J., Drummond, R., Engelhart, S. E., Kemp, A. C., et al. (2009). Holocene sea‐level changes along the
North Carolina Coastline and their implications for glacial isostatic adjustment models. Quaternary Science Reviews,28(17‐18),
1725–1736.
Hu, K., Ding, P., Wang, Z., & Yang, S. (2009). A 2D/3D hydrodynamic and sediment transport model for the Yangtze Estuary, China.
Journal of Marine Systems,77(1‐2), 114–136.
Hughes, S. A. (2002). Equilibrium cross sectional area at tidal inlets. Journal of Coastal Research,18, 160–174.
Jarrett, J.T. (1976). Tidal prism‐inlet area relationships, GITI Rep. No. 3, U.S. Army Engineer Waterways Experiment Station, Vicksburg,
MS.
Jevrejeeva, S., Moore, J. C., & Grinsted, A. (2010). How will sea level respond to changes in natural and anthropogenic forcings by 2100?
Geophysical Research Letters,37, L07703. https://doi.org/10.1029/2010GL042947
Jevrejeeva, S., Moore, J. C., & Grinsted, A. (2012). Sea level projections to AD2500 with a new generation of climate change scenarios.
Global and Planetary Change,80–81,14–20.
Jia, P., & Li, M. (2012). Circulation dynamics and salt balance in a lagoonal estuary. Journal of Geophysical Research,117, C01003. https://
doi.org/10.1029/2011JC007124
Kemp, A. C., Horton, B. P., Culver, S. J., Corbett, D. R., van de Plassche, O., Gehrels, W. R., et al. (2009). Timing and magnitude of recent
accelerated sea‐level rise (North Carolina, United States). Geology,37(11), 1035–1038.
Kemp, A. C., Horton, B. P., Donnelly, J. P., Mann, M. E., Vermeer, M., & Rahmstorf, S. (2011). Climate related sea‐level variations over the
past two millennia. Proceedings of the National Academy of Sciences,108, 11,017–11,022.
Kemp, A. C., Kegel, J. J., Culver, S. J., Barber, D. C., Mallinson, D. J., Leorri, E., et al. (2017). Extended late Holocene relative sea‐level
histories for North Carolina, USA. Quaternary Science Reviews,160,13–30.
Kopp, R. E., Horton, B. P., Kemp, A. C., & Tebaldi, C. (2015). Past and future sea‐level rise along the coast of North Carolina, USA. Climatic
Change,132(4), 693–707.
Leorri, E., Mulligan, R. P., Mallinson, D., & Cearreta, A. (2011). Sea‐level rise and local tidal range changes in coastal embayments: An
added complexity in developing reliable sea‐level index points. Revista de Gestão Costeira Integrada‐Journal of Integrated Coastal Zone
Management,11(3), 307–314.
Lesser, G. R., Roelvink, J. A., van Kester, J. A. T. M., & Stelling, G. S. (2004). Development and validation of a three‐dimensional mor-
phological model. Coastal Engineering,51(8‐9), 883–915.
Luettich, R. A. Jr., Carr, S. D., Reynolds‐Fleming, J. V., Fulcher, C. W., & McNinch, J. E. (2002). Semi‐diurnal seiching in a shallow, micro‐
tidal lagoonal estuary. Continental Shelf Research,22(11‐13), 1669–1681.
Mallinson, D., Culver, S., Leorri, E., Mitra, S., Mulligan, R., & Riggs, S. (2018). Barrier island and estuary co‐evolution in response to
Holocene climate and sea‐level change: Pamlico Sound and the Outer Banks Barrier Islands, North Carolina, USA. In L. Moore & B.
Murray (Eds.), Barrier Dynamics and the Impact of Climate Change on Barrier Evolution (pp. 91–120). New York: Springer Publishing.
https://doi.org/10.1007/978‐3‐319‐68086‐6_3
Mallinson, D. J., Smith, C. W., Culver, S. J., Riggs, S. R., & Ames, D. (2010). Geological characteristics and spatial distribution of paleo‐inlet
channels beneath the Outer Banks barrier islands, North Carolina, USA. Estuarine, Coastal and Shelf Science,88(2), 175–189.
Mallinson, D. J., Smith, C. W., Mahan, S., Culver, S. J., & McDowell, K. (2011). Barrier island response to late Holocene climate events,
North Carolina, USA. Quaternary Research,76(1), 46–57. https://doi.org/10.1016/j.yqres.2011.05.001
Mann, M. E., Woodruff, J. D., Donnelly, J. P., & Zhang, Z. (2009). Atlantic hurricanes and climate over the past 1,500 years. Nature,460,
880–883.
McGranahan, G., Balk, D., & Anderson, B. (2007). The rising tide: Assessing the risks of climate change and human settlements in low
elevation coastal zones. Environment and Urbanization,19(1), 17–37.
Metger, L. (2009). Holocene paleoenvironmental change in southern Pamlico Sound. Master's Thesis, East Carolina University, North
Carolina, pp. 1–178.
Moran, K. L., Mallinson, D. J., Culver, S. J., Leorri, E., & Mulligan, R. P. (2014). Late Holocene evolution of Currituck Sound, North
Carolina, USA: Environmental change driven by sea‐level rise, storms and barrier island morphology. Journal of Coastal Research.
https://doi.org/10.2112/JCOASTRES‐D‐14‐00069.1
Mulligan, R. P., Bowen, A. J., Hay, A. E., van der Westhuysen, A. J., & Battjes, J. A. (2008). Whitecapping and wave field evolution in a
coastal bay. Journal of Geophysical Research,113, C03008. https://doi.org/10.1029/2007JC004382
Mulligan, R. P., Hay, A. E., & Bowen, A. J. (2008). Wave‐driven circulation in a coastal bay during the landfall of a hurricane. Journal of
Geophysical Research,113, C05026. https://doi.org/10.1029/2007JC004500
Mulligan, R. P., Walsh, J. P., & Wadman, H. M. (2015). Storm surge and surface waves in a shallow lagoonal estuary during the crossing of a
hurricane. Journal of Waterway, Port, Coastal, and Ocean Engineering,141(4), A5014001. https://doi.org/10.1061/(AS CE)WW.1943‐
5460.0000260
Nichols, C.R., & Pietrafesa, L.J. (1997). Oregon inlet: Hydrodynamics, volumetric flux and implications for larval fish transport.
Department of Energy report DOE/ER/61425 –T3. 46 pp.
Passeri, D. L., Bilskie, M. V., Plant, N. G., Long, J. W., & Hagen, S. C. (2018). Dynamic modeling of barrier island response to hurricane
storm surge under future sea level rise. Climatic Change,149(3‐4), 413–425.
Passeri, D. L., Hagen, S. C., Medeiros, S. C., Bilskie, M. V., Alizad, K., & Wang, D. (2015). The dynamic effects of sea level rise on low‐
gradient coastal landscapes: A review. Earth's Future,3, 159–181. https://doi.org/10.1002/2015EF000298
Passeri, D. L., Hagen, S. C., Plant, N. G., Bilskie, M. V., Medeiros, S. C., & Alizad, K. (2016). Tidal hydrodynamics under future sea level rise
and coastal morphology in the Northern Gulf of Mexico. Earth's Future,4, 159–176. https://doi.org/10.1002/2015EF000332
Peek, K. M., Mallinson, D. J., Culver, S. J., & Mahan, S. A. (2014). Holocene geologic development of the Cape Hatteras Region, Outer
Banks, North Carolina, USA. Journal of Coastal Research,293(1), 41–58.
Purkiani, K., Becherer, J., Flöser, G., Gräwe, U., Mohrholz, V., Schuttelaars, H. M., & Burchard, H. (2015). Numerical analysis of stratifi-
cation and destratification processes in a tidally energetic inlet with an ebb tidal delta. Journal of Geophysical Research: Oceans,120,
225–243. https://doi.org/10.1002/2014JC010325
10.1029/2018JC014732
Journal of Geophysical Research: Oceans
MULLIGAN ET AL. 9256

Rahmstorf, S., Perrette, M., & Vermeer, M. (2012). Testing the robustness of semi‐empirical sea level projections. Climate Dynamics,
39(3‐4), 861–875.
Ranasinghe, R., Swinkels, C., Luijendijk, A., Roelvink, D., Bosboom, J., Stive, M., & Walstra, D. (2011). Morphodynamic upscaling with the
MORFAC approach: Dependencies and sensitivities. Coastal Engineering,58(8), 806–811.
Riggs, S.R., Ames, D.V. (2003). Drowning of North Carolina: Sea‐level rise and estuarine dynamics. University of North Carolina Sea Grant
Publication, Raleigh, NC, UNC‐SG‐03‐04. 152 p.
Riggs, S.R., Rudolph, G.L., Ames, D.W. (2000). Erosional scour and geologic evolution of Croatan Sound, Northeastern North Carolina.
North Carolina Department of Transportation, Raleigh, NC, Rept. No. FHWA/NC/2000‐02, 115 p.
Scavia, D., Field, J. C., Boesch, D. F., Buddemeier, R. W., Burkett, V., Cayan, D. R., et al. (2002). Climate change impacts on U.S. coastal and
marine ecosystems. Estuaries,25(2), 149–164.
Smith, E. (1999). Atlantic and East Coast Hurricanes 1900–98: A frequency and intensity study for the twenty‐first century. Bulletin of the
American Meteorological Society,80(12), 2717–2720. https://doi.org/10.1175/1520‐0477(1999)080<2717:AAECHA>2.0.CO;2
van der Westhuysen, A. J., Zijlema, M., & Battjes, J. A. (2007). Nonlinear saturation‐based white capping dissipation in SWAN for deep and
shallow water. Coastal Engineering,54(2), 151–170.
Wang, Z. B., Elias, E. P., van der Spek, A. J., & Lodder, Q. J. (2018). Sediment budget and morphologica l development of the Dutch Wadden
Sea: Impact of accelerated sea‐level rise and subsidence until 2100. Netherlands Journal of Geosciences,97(3), 183–214.
Wells, J. T., & Kim, S.‐Y. (1989). Sedimentation in the Albemarle‐Pamlico lagoonal system: Synthesis and hypothesis. Marine Geology,88,
263–284.
Zaremba, N., Mallinson, D., Leorri, E., Culver, S., Riggs, S., Mulligan, R. P., et al. (2016). Controls on the stratigraphic evolution and
paleoenvironmental change within a Holocene estuarine system: Pamlico Sound North Carolina, USA. Marine Geology,379, 109–123.
10.1029/2018JC014732
Journal of Geophysical Research: Oceans
MULLIGAN ET AL. 9257