Jcoastres D 14 00146.1

Influence of the Remote Forcing and Local Winds on the Barotropic
Hydrodynamics of an Elongated Coastal Lagoon

Author(s): Rafael Casares-Salazar and Ismael Mariño-Tapia
Source: Journal of Coastal Research, 32(1):116-130.
Published By: Coastal Education and Research Foundation
https://doi.org/10.2112/JCOASTRES-D-14-00146.1
URL: http://www.bioone.org/doi/full/10.2112/JCOASTRES-D-14-00146.1
BioOne (www.bioone.org) is a nonprofit, online aggregation of core research in the biological, ecological, and
environmental sciences. BioOne provides a sustainable online platform for over 170 journals and books published
by nonprofit societies, associations, museums, institutions, and presses.
Your use of this PDF, the BioOne Web site, and all posted and associated content indicates your acceptance of
BioOne’s Terms of Use, available at www.bioone.org/page/terms_of_use.
Usage of BioOne content is strictly limited to personal, educational, and non-commercial use. Commercial inquiries
or rights and permissions requests should be directed to the individual publisher as copyright holder.
BioOne sees sustainable scholarly publishing as an inherently collaborative enterprise connecting authors, nonprofit publishers, academic institutions, research
libraries, and research funders in the common goal of maximizing access to critical research.
Journal of Coastal Research 32 1 116–130 Coconut Creek, Florida January 2016
Influence of the Remote Forcing and Local Winds on the

Barotropic Hydrodynamics of an Elongated Coastal Lagoon
Rafael Casares-Salazar* and Ismael Mariño-Tapia
Department of Marine Resources
Centro de Investigación y de Estudios Avanzados del Instituto Politécnico Nacional (Cinvestav-IPN)
Mérida, Yucatán 97310, México
ABSTRACT
Casares-Salazar, R. and Mariño-Tapia, I., 2016. Influence of the remote forcing and local winds on the barotropic
hydrodynamics of an elongated coastal lagoon. Journal of Coastal Research, 32(1), 116–130. Coconut Creek (Florida),
ISSN 0749-0208.
The main objective of this study is to investigate the effects of local winds and sea-level variations at tidal and subtidal
frequencies on the water levels, depth-averaged velocities, and volume fluxes of an elongated and shallow coastal lagoon
using field measurements and a barotropic two-dimensional numerical model. Tides are mainly diurnal and attenuate
~85% from the mouth to the lagoon head, indicating a barotropic propagation dominated by friction. On the other hand,
the low-frequency water-level fluctuations propagate with little attenuation (~17%), dominating the water-level
variability at the lagoon head. It is considered important to elucidate whether this low-frequency variability is related to
local winds or to remote effects. Modeling results show that local winds can increase water levels up to ~1.9 cm in the
lagoon head (5% of tide level), whereas remote sea-level variability accounts for up to ~26.6 cm (320% of astronomical
tidal level), clearly dominating lagoon water levels. On the other hand, diurnal sea breezes play a minor role in water
levels inside the lagoon and do not reinforce diurnal tides in order to generate an important set-up. The cumulative water
volume is highly correlated with sea level and shows that low-frequency sea-level fluctuations are especially efficient at
accumulating or exporting water from the lagoon during periods of ascending or descending sea level, respectively. This
implies that water exchange between the lagoon and the coastal ocean is dominated not only by diurnal tides but also by
the low-frequency sea-level fluctuations of astronomical and meteorological origin (i.e. wind forcing through Ekman
transport).
ADDITIONAL INDEX WORDS: Low-frequency sea-level fluctuations, wind forcing, tide propagation, numerical
modeling.
INTRODUCTION Although tides can be important drivers for the dynamics of

Coastal lagoons provide important ecosystem services be- coastal water bodies, the low-frequency variability may be
cause they are nursery grounds for various marine species, particularly significant because tides may be heavily attenuat-
natural sources of seafood and sea salt, and often important ed within the lagoon, and the lower frequency fluctuations could
sites for aquaculture. Furthermore, they absorb storm energy, be of comparable or greater magnitude (Wong, 1987). Low-
provide flood protection and drainage, and are safe ports for frequency variability can be separated into local and remote
fishermen and industry. Their landscapes also have an components. Local winds act directly over the surface area of an
important aesthetic value, making them popular sites for estuary, modifying the residual currents as a result of an along-
tourism and recreation as well as for human habitation. estuary wind stress (Wong and Valle-Levinson, 2002; Wong and
As coastal lagoons contribute significantly to economic, Wilson, 1984). On the other hand, along-shore remote winds
ecological, and social welfare, it is important to understand cause changes in the sea level attributable to Ekman transport,
the processes that play a dominant role in the exchange of affecting several hundreds of km of coast and generating
pollutants, organisms, salt, heat, and sediments between these coastal-trapped waves (Santoro, Fossati, and Piedra-Cueva,
water bodies and the coastal ocean. The hydrodynamics of 2013; Smith, 1978; Wang and Elliott, 1978; Wong and Moses-
estuaries and coastal lagoons is driven by tides, wind stress, Hall, 1998). These oceanic features can cause the sea-level
river inputs, and heat fluxes, which act at different time scales. fluctuations at the estuary mouths that propagate into the
In this regard, numerical models have been a convenient tool systems with little or no attenuation (Möller et al., 2001; Shetye
for understanding and evaluating the influence of the various and Vijith, 2013; Smith, 1977; Wong, 1987; Wong and Valle-
physical forcings on the circulation and exchange processes in Levinson, 2002), influencing water exchanges between the
these water bodies (e.g., De Marchis et al., 2012; Jia and Li, inlets and the ocean (Waterhouse et al., 2013; Wong and Moses-
2012; Malhadas et al., 2010; Sankaranarayanan and Fringer, Hall, 1998; Wong and Valle-Levinson, 2002). Several authors
2013; Wijeratne and Rydberg, 2007). have identified this remote effect to be the dominant factor
controlling the subtidal sea-level fluctuations in the interior of
DOI: 10.2112/JCOASTRES-D-14-00146.1 received 31 July 2014; several estuaries, with local winds playing a secondary role
accepted in revision 30 December 2014; corrected proofs received
(Garvine, 1985; Janzen and Wong, 1998; Wong, 1987).
12 March 2015; published pre-print online 20 April 2015.
*Corresponding author: [email protected] Nevertheless, the barotropic exchange of volume and mass
Ó
Coastal Education and Research Foundation, Inc. 2016 can also be influenced by local winds as they generate
Remote and Local Forcing at a Coastal Lagoon 117
important gradients along the water surface, affecting currents

within the lagoon. For instance, Sankaranarayanan and
Fringer (2013) found important low-frequency elevations
attributable to remote effects in San Francisco Bay, but the
low-frequency currents were more influenced by local winds, as
suggested from numerical experiments forced with steady
winds. Local winds blowing toward the head of an estuary or
coastal lagoon drive a two-layer circulation with surface inflow
and a bottom return flow as a result of the water-level set-up
inside the basin, as observed by several authors (e.g., De
Marchis et al., 2012; Iglesias and Carballo, 2010; Waterhouse et
al., 2013; Wong and Moses-Hall, 1998); this two-layer circula-
tion superimposes on the unidirectional effect of the remote
Figure 1. Study area map and observation sites.
forcing (Wong and Valle-Levinson, 2002).
For extraordinarily large and choked systems, such as the
Patos lagoon in Brazil, subtidal sea-level oscillations at the
ocean boundary are confined to the lower estuary and are fluctuation of water levels within the lagoon, followed by
damped in its interior. Further inland, the local forcing several aspects of tidal propagation, as well as a number of
generated by winds or freshwater inputs is likely to be the water exchange and modeling implications of this study.
main forcing controlling the dynamics of the system (Fer- Finally, conclusions are drawn.
nandes et al., 2004).
The relative importance of local winds vs. remote forcing on Description of the Study Area
the low-frequency dynamics of coastal lagoons is bound to The study site is Celestun, a tropical coastal lagoon located in
depend on the geometry of the lagoon, the position and number the NW of the Yucatan Peninsula, Mexico (20845 0 –20858 0 N and
of mouths, the local wind direction (Mied et al., 2010), and the 90819 0 –90825 0 W). The coastal lagoons of Yucatan are depres-
characteristic frequency and intensity of the wind events, sions roughly parallel to the coast, separated from the sea by a
which have a seasonal variability (Wong and Valle-Levinson, barrier island comprising biogenic sand where wave action and
2002). No single cause-and-effect behavior is evident (Mied et coastal currents deposit the sediments that contribute to its
al., 2010) because processes occur synchronously; therefore, formation (Lankford, 1977). The coastal lagoon of Celestun is
more contributions on this topic are necessary to better around 22 km long, with a permanent inlet that is 650 m wide
understand the exchange processes between coastal lagoons at its southern end (Figure 1). It has a variable width ranging
and the ocean. from 300 m near its middle zone up to 2.2 km at the head. The
The objective of this study is to analyze the influence of subtidal water body covers a surface of around 25.16 km2. In
water-level variations on the hydrodynamics and exchange the middle region there is an embankment road of 395 m that
processes of an elongated and shallow coastal lagoon, focusing obstructs the circulation and restricts the water flow through a
on the effects of low-frequency fluctuations (which account for reduced cross-section of around 140 m under a bridge, set on
the remote and local forcing), but also including the effects of round concrete piles (see Figure 1). The bridge and the
higher frequencies (sea breezes and astronomical tides). embankment road, built in 1979, modified the tidal channel
Because of the elongated characteristics of the study site and trajectory as well as the sedimentation on both sides of the
the presence of only one mouth at one end, local winds are road. The lagoon is a shallow system with a variable depth
expected to play an important role in modifying the water ranging from nearly 0.5 m at the tidal flats to ~2 m in the main
levels, especially at the head. The main tool for this study is a tidal channel, which meanders along the southern half of the
two-dimensional (2D) (depth-averaged) numerical model lagoon from the mouth up to 3 km north of the bridge with a
(Delft3D), implemented to determine the relative importance mean width of around 120 m; the last 9 km toward the head has
of the aforementioned processes on the barotropic dynamics of no notorious channel. The adjacent marine area is shallow with
the lagoon. depths of 3.9 m that is 5 km away from the inlet in the offshore
The paper is structured as follows. The next subsection direction (Figure 2).
describes the study area. This is followed by the methodology, Along the lagoon coastline, there is a mangrove fringe around
which deals with the field campaign, the data treatment, the 150 m wide (Sánchez-Arguelles, 1994) comprising Rhizophora
numerical model set-up, and its verification. In the ‘‘Results’’ mangle, Avicennia germinans, Laguncularia racemosa, and
section, we present the field data for the water levels and Conocarpus erectus (Herrera-Silveira, 1996). North and east of
longitudinal velocities, the attenuation of the tidal wave that the lagoon, the mangrove wetland extends for several km. The
propagates toward the lagoon head, and the behavior of the lagoon has high biodiversity and is an essential nursery area
remotely originated low-frequency fluctuation, including a for the American flamingo Phoenicopterus ruber ruber (Barri-
long-term analysis of its persistence. The modeling results os-Espino, 1988). The local economy relies on the use of natural
include the effects of the different physical forcings on the resources by artisanal fisheries, which capture shrimp juve-
water levels, depth-averaged longitudinal velocities, volume niles and fish, and by ecotourism activities (Pech, Ardisson, and
fluxes, and cumulated volumes in the lagoon. We then discuss Hernández-Guevara, 2007). This lagoon is a protected natural
the role of the remote and local winds on the low-frequency area denominated Biosphere Reserve of Ria Celestun. It is also
Journal of Coastal Research, Vol. 32, No. 1, 2016

118 Casares-Salazar and Mariño-Tapia
Figure 3. Temperature and salinity profiles in the lagoon.
fresher water at the head (21 psu) and saltier water at the
mouth (37 psu). Despite this marked horizontal salinity
difference, these profiles clearly show that the water column
was well mixed with no evident stratification.
Figure 2. (a) Model domain with the bathymetry and open boundaries; (b)
METHODS
grid resolution. (Color for this figure is available in the online version of this
paper.) This section describes the methods used to obtain the lagoon
bathymetry and to monitor the hydrodynamic variables along
with the techniques used for processing the data. It also
describes the numerical model set-up, performance parame-
a RAMSAR site (see Ramsar, 2015) of international importance ters, and analysis methods used to calculate water-level
for the conservation and responsible use of wetlands and their attenuations and time lags from the mouth to the lagoon head.
resources.
There are three typical climatological seasons in the area Bathymetric Survey
(SARH, 1989): the dry season is from March through May (0–50 The bathymetric survey of the lagoon and adjacent marine
mm rainfall); the rainy season is from June through October area (Figure 2) was carried out in a boat equipped with an echo-
(.500 mm); and the winter storms (northerlies) occur from sounder (SyQwest Bathy500-DF) coupled with a differential
November through February (20–60 mm) and are character- GPS (DGPS, Leica GPS System 1200) to correct the recorded
ized by strong northerly winds greater than 80 km/h and low depths for local water-level fluctuations. The shallower and
air temperatures (,228C). The northerlies are imposed by polar narrow zones were surveyed with a jet ski. Several topographic
cold fronts traveling southward. Mean annual rainfall in the profiles were carried out at the intertidal zones using a DGPS
Celestun area is 747 mm (Herrera-Silveira, 1994), whereas the mounted on a mobile aluminum cart. The topographic and
evaporation rate is ~1400 mm (Herrera-Silveira et al., 1999). bathymetric levels were referenced to the mean sea level (msl)
This excess evaporation could imply a reduction in lagoon area, for the study area based on a year of tidal records at Tide Gauge
especially during the dry season; nevertheless, Celestun has a (TG) station (Figure 1) located at Celestun pier; the sampling
permanent input of groundwater that can replace the water interval for the sea level at TG was 6 minutes. The bathymetry
loss attributable to evaporation. On the other hand, the was spatially interpolated in GSþ (Gamma Design Software)
permanent communication with the ocean through the inlet version 9.0. An ordinary Kriging was applied using a Gaussian
also replaces water lost through evaporation. The Yucatan isotropic variogram with a distance interval of 20 m.
Peninsula is a karstic region where the high permeability of the Field Campaign and Data Treatment
bedrock prevents the formation of rivers and reduces surface A set of oceanographic instruments were deployed along the
runoff (Perry, Velazquez-Oliman, and Marin, 2002). Precipita- lagoon to monitor hydrodynamic variables, that is, water
tion rapidly penetrates the ground and is incorporated into the salinity, water levels, and flow velocities (Figure 1). This
subterranean aquifer. Groundwater discharges from the deployment consisted of an acoustic velocimeter (Nortek-
Yucatan aquifer have been detected in many places along the Vector) at observation site 1 (S1); a CTD diver (Schlumberger)
coast. This includes the Celestun region, where at least 30 at observation site 2 (S2); a multiparametric probe (YSI) at
submarine springs have been identified throughout the lagoon, observation site 3 (S3); an acoustic current-profiler (Nortek-
with the majority concentrated in the middle and northern Awac) at observation site 4 (S4); a CTD diver (Schlumberger) at
zones. The discharge of fresh groundwater from these springs observation site 5 (S5); an acoustic velocimeter (Nortek-Vector)
creates a strong salinity gradient, with values from 5–21 psu in and a multiparametric probe (YSI) at observation site 6 (S6);
the lagoon interior to 33–38 psu at its mouth showing seasonal and an acoustic current profiler (SonTek-Argonaut) along with
and interannual variability. Several salinity and temperature a CTD (Sea-Bird Electronics) at the observation site denomi-
profiles were carried out during the field campaign of this study nated ‘‘Sea,’’ located approximately 5 km offshore from the
(Figure 3) using a CTD (an oceanographic instrument that mouth. The acoustic velocimeter at S1 was installed vertically
registers conductivity, temperature and depth), which showed in a very shallow region; therefore, the water level was


occasionally below the pressure sensor interrupting the signal SD2
a¼ 1 3 100; ð1Þ
record. On the other hand, the velocity probe was continuously SD1
submerged, registering velocity at all times. S1 and S2 were
where SD1 is the signal standard deviation at the upstream
located on tidal flats at the head of the lagoon because no main
station and SD2 is the signal standard deviation at the
channel is present there; S3 was located by the lagoon margin
downstream station. This method is only applicable when the
for security reasons; S4, S5, and S6 were located in the main
signals are symmetric around their mean (i.e. there is no
channel. These instruments recorded one data point every 15
tendency) and when the readings are continuous (Erskine,
minutes from 24 May to 9 July 2012 (46 days). The sampling
1991). To calculate the attenuations for the tidal (diurnal)
interval is considered short enough to solve all the processes of signal, Eq. 1 was applied to a detrended section of 3 days (15–18
interest to this study (semidiurnal and lower frequencies), and June). Time lags were calculated by applying a cross-
processes at higher frequencies are excluded because they are correlation for the same data period. Alternatively, the
not the focus of this investigation. attenuations were calculated with the square roots of the
Meteorological records (i.e. wind velocity and direction) power spectral densities (PSDs) instead of the SDs in Eq. 1,
sampled every 10 minutes were obtained from a weather obtained from a spectral analysis of all the time series (43 days)
station (Davis Vantage Pro2) located at a height of 10 m in the using the Welch periodogram method with a Hanning window
middle zone of the lagoon, next to the bridge (WS in Figure 1). of 1024 data points. This included four nonoverlapping
These records are assumed to be representative of the entire intervals to provide a level of statistical confidence.
system because there are no important topographic features in
the region. Wind direction is given following meteorological Numerical Model Set-up
The numerical model used was Delft3D (Deltares Systems).
convention, that is, from where wind blows, with 08 represent-
This model is capable of calculating the nonsteady free-surface,
ing north and advancing clockwise up to 3608. A barodiver
flow, and transport phenomena resulting from tidal and
(Schlumberger) that registers the barometric pressure was also
meteorological forcing in two (2DH, depth-averaged) or three
placed at WS to correct the pressure readings from the CTD
dimensions (3D). The system of governing equations consists of
divers because they measure the absolute pressure. The
the horizontal momentum, continuity (or mass conservation),
ellipsoidal height for the pressure sensor at the Sea station
and transport equations under the shallow water (hydrostatic
was obtained from a differential GPS while navigating over the
pressure relation) and Boussinesq (water incompressibility)
instrument during the bathymetric survey, so this pressure
assumptions. These equations are solved at each time step in a
sensor and its corresponding readings were referenced to the
finite differences curvilinear boundary-fitted grid, which
msl. For the rest of the water-level time series inside the results in the prediction of coastal hydrodynamics (Lesser et
lagoon, the arithmetic means were subtracted from the al., 2004). A 2DH model was set up to study the barotropic
corresponding signals to obtain the time series with a zero hydrodynamics of a coastal lagoon, with emphasis on water
mean level as a first approximation to assess the water-level levels, volume fluxes, and cumulative volumes through cross-
variability. sections. This represents a first approximation of the hydrody-
Measured and simulated velocity components (X and Y) were namics in the system. Considering that the lagoon is very
rotated at each observation site to align the Y velocity shallow and that the water column was well mixed during the
component with the main channel or the longitudinal axis of period of this study, with no important vertical gradients in
the lagoon, with positive directions toward the head. Rotation density (see Figure 3), the use of a depth-averaged 2D model is
angles were obtained based on the minimization of the root justified. What this model cannot solve is the detailed behavior
mean square error (RMSE) of the rotated X velocity component, of the vertical velocity field, which has been studied even in
thus the rotated Y velocity component was maximized. shallow lagoons (e.g., De Marchis et al., 2012; Janzen and
Measured time series (stations S1–S6 and Sea) were processed Wong, 1998). This topic, although interesting, would require a
to remove spurious data that lie outside the average 63 3D model, which escapes the scope of the present study.
standard deviations (SDs) and were replaced with the average The model domain was defined up to the continental limit of
value of the neighbors. To focus on the characteristic frequency the mangrove fringe, which is considered to be flooded several
range of the processes of this study, high frequencies (i.e. waves days of the year. This limit was digitized in Envi 4.7 (Exelis)
and other processes) were filtered with an adaptive detrending from a SPOT-5 satellite image with a spatial resolution of 10 m
algorithm (Hu, Gao, and Wang, 2009). This procedure consisted and four spectral bands. An adaptive grid of 1730 3 291 nodes
of fitting a third-order polynomial to segments of 31 data points was created in Cartesian coordinates. The grid lines were
(7.75 hours). This algorithm effectively eliminates any jumps or traced following the main channel and the lagoon longitudinal
discontinuities around the boundaries of neighboring segments axis, where the preferential flow direction would be; thus, the
and is more accurate than a moving-average filter. horizontal resolution is variable from 8 to 26 m, which
appropriately solves the bathymetric gradients and the
Water Level Attenuations and Time Lags geometrical shape of the lagoon (Figure 2). The model supports
Signal attenuations (a) for water levels propagating toward the wetting and drying of the grid cells to take into account the
the head of the lagoon were calculated as the quotients of their water-level fluctuations in the intertidal region. Model obser-
SDs at different sites based on the concept of tide-transmission vation points were defined at the same sites where oceano-
efficiency, defined by Erskine (1991). Here, we used the graphic instruments were deployed in the field. Flow cross-
expression sections were also defined at the bridge and at the mouth of the

Table 1. Errors for the different Manning roughness coefficients during the calibration process.
Manning Roughness Coefficients (n)

Mean RMSE Mean REV
Model (Preliminary) Main Channel Tidal Flats Head Water Levels (m) (Dimensionless)
62 0.012 0.05 0.05 0.072 0.435
66 0.018 0.05 0.05 0.065 0.383
68 0.018 0.07 0.07 0.05 0.372
69 0.018 0.07 0.11 0.05 0.372
lagoon. The bridge piles were modeled as subgrid structures boundaries (Figure 2). Six simulations or model tests (MT)
that account for friction losses, whereas the embankment road were conducted. A summary of the numerical experiments and
on the landward side, prior to the bridge, was defined as dry their boundary conditions is given in Table 2. MT-1 is a special
cells. simulation in which only wind forcing is included at the surface
At the surface, the kinematic boundary condition was boundary in order to ignore other forcings and assess the effects
imposed. The free surface at this boundary was forced with of winds on the barotropic dynamics. In this model, the offshore
temporally variable but spatially constant wind measure- and cross-shore open boundaries were prescribed with a
ments. Wind effects were transformed to shear stress values constant water level of zero (i.e. m above msl) during all of
(ss) with the widely used quadratic friction relation (Gerritsen the simulation but were allowed the flux of momentum through
et al., 2007), the open boundaries. MT-2–MT-5 were run to verify the
2
numerical model. For this purpose, MT-2 and MT-3 were both
jss j ¼ qa Cd U10 ; ð2Þ forced without and with local winds, respectively; in both
where qa is the air density, Cd is the wind drag coefficient, and models the offshore open boundary was prescribed with water-
U10 is the measured wind velocity at 10 m height (WS in Figure level time series of astronomical tides (all the tidal constitu-
1). The air density was set to 1.15 kg/m3, and the wind drag ents), whereas the cross-shore open boundaries were forced
coefficient was set to 0.0013, which is suggested as an with along-shelf vertically averaged velocities measured at Sea
appropriate value for engineering calculations (Fischer et al., station (with the SonTek-Argonaut). The astronomical tide was
1979). A free-slip boundary condition was defined along all obtained from a harmonic analysis applying T_TIDE (Pawlo-
closed lateral boundaries. This boundary condition has been wicz, Beardsley, and Lentz, 2002) with a 43-day time series of
previously used to study the effects of winds on the circulation the sea level measured at Sea station (with the CTD Sea-Bird
of Ria de Muros, Spain (Iglesias and Carballo, 2010); in the Electronics). Similarly, MT-4 and MT-5 were forced without
present study, this does not seem to substantially affect model and with local winds, respectively. In both models, the offshore
performance because the model was successfully verified. On open boundary was forced with measured sea-level data at Sea
the other hand, the definition of the bottom boundary condition station, thus considering the remote forcing; the same along-
was crucial to achieve an appropriate verification of the model. shelf velocities of the previous models (MT-2 and MT-3) were
Through field observations of bed characteristics and a series of prescribed at the cross-shore open boundaries. MT-1–MT-5
preliminary numerical sensitivity tests (Table 1), a detailed were run for a period of 15 days. MT-6 is a long-term run
spatially variable bed roughness distribution was prescribed. performed to corroborate the effects of the remote forcing on the
The variable friction was set in the model using Manning water-volume exchanges between the lagoon and the ocean. A
coefficients (n) with values in agreement with those reported by 43-day time series could have limitations to properly analyze
Chow (1959). A coefficient of 0.018 was assigned to the marine the effects of subtidal variability inside the lagoon; hence, this
zone and along the main channel given their greater depth and MT was forced at the offshore open boundary with 2 months’
consequently lower energy losses; a coefficient of 0.07 was worth of sea-level data measured at the TG station. The local
assigned to the tidal flats and intertidal zones; and a coefficient wind was not included in this simulation because only the
of 0.11 was assigned to the head of the lagoon by virtue of the effects of the remote forcing were evaluated. No information on
greater submerged vegetation density, which occupies almost along-shelf currents was available for the same period,
the entire water column. therefore, a Neumann condition (along-shore water-level
The computations were executed using a constant eddy gradient) at the cross-shore open boundaries was prescribed.
viscosity of 1 m2/s. Since this parameter depends on the flow Roelvink and Walstra (2004) document that a Neumann open
characteristics and grid size (Deltares, 2011), a sensitivity boundary condition can be used instead of prescribing a fixed
analysis was carried out using two more values (0.3 and 3 m2/s). water level or velocity, allowing the model to determine the
The resultant water levels and velocities were very similar for correct solution. In many cases, this parameter can be assumed
these three tests; therefore, the value of 1 m2/s was chosen to be zero: only in tidal cases where a storm surge travels along
based on numerical stability and running time criteria. a coast does the along-shore gradient vary in time (Roelvink
Furthermore, this value has been used by other authors in and Walstra, 2004). During some preliminary tests (not
hydrodynamic numerical simulations (e.g., Luijendijk, 2001; presented in this paper), models forced with measured along-
Roelvink and Walstra, 2004). The time step for all simulations shore velocities or with the Neumann condition (with a zero
was set to 0.5 minutes for numerical stability reasons. along-shore water-level gradient) at the cross-shore open
Three open boundaries were defined at the adjacent marine boundaries gave the same results for water-level and velocity
zone: one offshore, parallel to the coast, and two cross-shore time series inside the lagoon. Noticeable differences were only

Table 2. Numerical experiments summary.
Open Boundary Conditions

Offshore Boundary Cross-Shore Boundaries
Model Boundary Type (Value) Boundary Type (Value) Wind Forcing Simulation Period Duration Purpose of the Model
MT-1 WL (zero, m amsl) WL (zero, m amsl) Yes 15–30 June 2012 15 days
Isolate the effects of the
local wind
MT-2 WL (TS of astronomical tides) Current (along-shelf velocities) No 15–30 June 2012 15 days Verification, only
astronomical tides
MT-3 WL (TS of astronomical tides) Current (along-shelf velocities) Yes 15–30 June 2012 15 days Verification,
astronomical tides and
winds
MT-4 WL (TS of SL, measured at Sea) Current (along-shelf velocities) No 15–30 June 2012 15 days Verification, remote
forcing and tides
MT-5 WL (TS of SL, measured at Sea) Current (along-shelf velocities) Yes 15–30 June 2012 15 days Verification, remote
forcing, tides and wind
MT-6 WL (TS of SL, measured at TG) Neumann No 17 Dec. 2010–16 Feb. 2011 2 months Corroborate the effects of
the remote forcing
(subtidal sea-level
fluctuations)
MT: model test; WL: water level; m amsl: meters above mean sea level; TS: time series; SL: sea level; Sea: Sea station; TG: Tide Gauge station.
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
found outside the system. Because the focus of this research is 1
on the processes inside the lagoon, the Neumann boundary RMSE ¼ RðOi Pi Þ2 : ð4Þ
n
condition with a zero water-level gradient is considered to be
suitable. The RMAE is useful for evaluating predicted velocities,
Summarizing, the field campaign data (S1–S6, Sea station, though Kankara, Mohan, and Venkatachalapathy (2013) also
and wind data at WS) covered the period from 24 May to 9 July used it to evaluate water levels. The RMAE is defined by
2012 (46 days). A subset of these data, from 15–30 June 2012 Walstra et al. (2001) as
(15 days), was used to force MT-1 to MT-5. Data from a previous hjVm Vc ji
period of sea-level measurements at TG (i.e. from March 2010 RMAE ¼ ; ð5Þ
hjVm ji
to August 2011 [17 months]) and wind data from November
2010 to August 2011 were used to further investigate the where Vm and Vc are the measured and computed velocity
persistency of the subtidal oscillations and its origin. The vectors, respectively. Angular brackets indicate an arithmetic
numerical test MT-6 was forced with a subset of these data mean. This statistical parameter is particularly useful for the
spanning 2 months from 17 December 2010 to 16 February comparison of current velocities since it accounts for both the
2011. MT-6 will be referred to as the long-term simulation. magnitude and direction of the flow. It is also called the REV,
documented by Van Rijn et al. (2002), and is used in other
Model Performance Parameters studies (e.g., Kankara, Mohan, and Venkatachalapathy, 2013;
Prior to using the model for numerical experimentation, it is Sousa and Dias, 2007; Walstra et al., 2003; Walstra, van Rijn,
necessary to verify that its results are a reasonable approxi- and van Helvert, 2002). Experience with this parameter is
mation of reality (Luijendijk, 2001; Walstra, van Rijn, and van limited, and an associated qualitative interpretation based on
Helvert, 2002). Therefore, modeled results were compared with Van Rijn et al. (2002) has been included in Table 3.
the measured data. To quantify the model uncertainty three Table 4 shows the error estimates between the verified model
parameters were used: the Nash-Sutcliffe model efficiency and the measured data at each observation site. Site 1 (S1) was
coefficient (E), the RMSE, and the relative mean absolute error excluded since flow velocity is very low, and water levels could
(RMAE), also called relative error vector (REV). not be completely measured as the instrument was exposed to
The Nash-Sutcliffe model efficiency coefficient E (Nash and the air for prolonged periods. Three out of five sites reached an
Sutcliffe, 1970) was calculated according to the following ‘‘excellent’’ rank for water levels (RMAE), whereas the rank for
equation: velocities (REV) was between ‘‘reasonable/fair’’ and ‘‘good.’’
RjOi Pi j2 Both variables had an average rank of ‘‘good.’’ The model
E¼1 ; ð3Þ efficiency coefficient (E) also presented high values, with an
RjOi Oj2
average of 0.857 for surface elevations and 0.901 for velocity.
where Oi and Pi are the observed and predicted values of the Another parameter that helps to assess model performance is
variable at each time step and O is the mean of the observed an analysis of the measured and simulated tidal attenuations
values of the variable. Perfect agreement is achieved for E ¼ 1. and time lags for the verified model MT-5 (Table 5). Only the
This parameter is a quantitative measure of the model’s diurnal frequencies are included. There is good correspondence
goodness of fit and has recently been used by De Marchis, between the measured and simulated data, with a consistent
Freni, and Napoli (2014). increment in the attenuation and time lags from the inlet
The RMSE represents the differences between the values toward the head. This result, along with the values obtained for
calculated by the model and those actually measured, E, RMAE, and REV, confirm that the 2D depth-averaged

Table 3. Classification of error ranges for RMAE or REV (according to Van and direction, respectively, measured at WS. It is interesting to
Rijn et al., 2002). note that moderate and persistent winds from SW (~2308)
around 24 June coincide with the water superelevation seen
Qualification REV
throughout the lagoon, especially at the head (S2). This
Excellent ,0.2
observation could suggest that local winds affect the water
Good 0.2–0.4
Reasonable/Fair 0.4–0.7 levels in the lagoon considerably. This issue is addressed in the
Poor 0.7–1.0 ‘‘Numerical Experiments’’ section.
Bad .1.0 Harmonic analysis of the water level measured at Sea and at
TG stations shows two main constituents for the astronomical
numerical model is capable of reproducing the hydrodynamic tide, both at diurnal frequencies (Table 6): the K1 and the O1,
behavior of the lagoon with good accuracy; therefore, the model the latter with a smaller amplitude. For the field measurement
is suitable for performing numerical experiments to investigate period, the astronomical tide has a range of around 1.05 m
the effect of the physical forcings (sea-level variations, during spring tides and 0.16 m during neap tides; the residual
astronomical tides, and local winds) on water levels and signal (i.e. measured sea level minus the astronomical tide)
longitudinal depth-averaged velocities in the coastal lagoon of reveals the low-frequency fluctuation more clearly. The
Celestun, Mexico. maximum negative deviation is 27.2 cm (below the msl) on 3
June (Figure 6), and the maximum positive is þ34.8 cm near 24
RESULTS June. The presence of such low-frequency behavior in sea level
The results for measured water levels and longitudinal at a distance of 5 km offshore of the lagoon mouth, that is, far
velocities are presented in this section, as well as for those enough from the lagoon influence, suggests that this phenom-
obtained with the numerical experiments, which include the enon could be more linked to remote effects.
model verification, the effects of local winds, and the remote The PSD function of the water level at Sea station (one
forcing on water levels and velocities in the lagoon and volume nonoverlapping interval, Figure 7a) clearly shows two peaks at
exchanges with the coastal ocean. the diurnal frequencies representing the K1 and O1 tidal
Because wind is one of the main forces under investigation in constituents, with K1 amplitudes greater than O1, consistent
this study, a wind rose for the field campaign and for the long- with the harmonic analysis results (Table 6). Two peaks for the
term data is presented in Figure 4a. During the 46 days of the semidiurnal frequencies are also shown, representing the M2
field campaign, the local wind was predominantly from the and S2 tidal constituents, with M2 amplitudes considerably
ENE and E (67.58 to 908), directions that would not be expected greater than S2, also in accordance with harmonic analysis.
to considerably affect the lagoon water levels. Nevertheless, the Finally, an important low-frequency constituent with a period
few events with a wind direction along the lagoon axis (toward of 21.33 days is also shown. The statistical confidence of Figure
the head) seem to coincide with low-frequency water-level
7a was corroborated by performing a PSD with shorter
variability (see Figure 5). In the longer term (Figure 4b), the
windows of 1024 data points or four nonoverlapping intervals.
local wind was predominantly from the E and ESE (908 to
The results are shown in Figure 7b.
112.58), which will also have little effect on the water levels.
Using the SDs calculated from the PSD of the different
Northerly winds (N and NNE) are not unimportant, and their
stations (Eq. 1), we can conclude that the signal at diurnal
direction (along axis) is such that it affects water levels in the
frequencies attenuates around 84.6% from the mouth (S6) to
lagoon and the coastal ocean. A section is dedicated to analyze
the head (S2) of the lagoon (Table 5), which is similar to the
the origin of these oscillations at the end of the ‘‘Results’’
attenuation calculated from the SDs derived from the time
section.
series for a 3-day data period (87.7%). This indicates a
Measured Water Levels barotropic tidal propagation toward the head of the lagoon
Water-level time series from the measured data (Figure 5a) with an important remaining tidal signal of around 15% at the
reveal a propagating tidal signal characterized by frictional head (S2) compared to that at the mouth (S6). This tidal
damping and a corresponding phase lag from the ocean to the (diurnal) signal takes around 7.75 hours to travel from S6 to S2,
lagoon head. The signal also shows a low-frequency (subtidal) a value estimated from a cross-correlation function of the
water-level oscillation with a slight water-level depression measured data. Similarly, as a first-order approximation, the
around 29 May and an increase near 24 June, noticeable at all low-frequency signal with a period of 21.33 days attenuates
stations including Sea. Figures 5b and 5c show the wind speed ~17% from S5 to S2; in other words, a subtidal water-level
Table 4. Error estimates between the verified model (MT-5) and measured data at each observation site.
Water Levels Velocities

Site E RMSE (m) RMAE RMAE (qualitative) E (for Y vel.) RMSE (for Y vel., m/s) REV REV (qualitative)
S2 0.8537 0.061 0.271 Good - - - -
S3 0.5279 0.070 0.687 Reasonable/Fair - - - -
S4 0.9734 0.036 0.116 Excellent 0.9136 0.0953 0.310 Good
S5 0.9386 0.061 0.199 Excellent - - - -
S6 0.9934 0.020 0.072 Excellent 0.8874 0.1302 0.433 Reasonable/Fair
Mean 0.8574 0.050 0.269 Good 0.901 0.113 0.372 Good

Table 5. Attenuations and time lags for measured and simulated water levels at diurnal frequencies, for the verified model (MT-5).
Measured Water Levels Simulated Water Levels

Sites (from/to) Attenuation with PSD (%) Attenuation with SD (%) Time Lag (h) Attenuation with SD (%) Time Lag (h)
S6–S5 14.4 11.3 0.75 8.9 0.50
S6–S4 30.5 28.0 1.75 32.1 1.75
S6–S3 - 72.5 3.50 54.4 3.00
S6–S2 84.6 87.7 7.75 80.9 8.25
S6–S1 - - - 81.1 8.50
PSD: power spectral density; SD: standard deviation. Blank: not available.
signal of ~83% remains at the head. This value was obtained Effects of Local Winds on Lagoon Water Levels and
using the PSD-derived SDs. Velocities
The effect of local winds on water levels was tested with the
Measured Longitudinal Velocities
numerical model forced just with winds and zero water levels at
Longitudinal (along-channel) velocity time series from
the open boundaries (MT-1). This was done to isolate all the
measured data reveal that velocities are almost imperceptible
other forces present in nature that affect the lagoon hydrody-
at the head (Figure 8). A cross-correlation function between S6
namics. This experiment reveals a greater set-up/set-down at
(mouth) data and S1 (head) has a very low correlation
the head of the lagoon with a slope toward the inlet, where the
coefficient (0.36), which indicates that the variability at the water level tends to zero. The maximum wind set-up/set-down
head is unrelated to that at the inlet. The very weak currents at the lagoon head was only 3.9 and 5.9 cm, respectively, for the
seem to be controlled by the high friction values and the closed realistic wind data used in the model (Figure 9). The maximum
nature of the system, which gradually reduces the fluxes from wind intensity peaks that caused the greater set-down (15 and
the mouth. Maximum tidal currents are observed at the mouth 19 June) had a persistent direction of around 508 (NE winds),
(S6, with a maximum value of 1.16 m/s) and tend to occur showing that these winds acted for enough time to force water
during spring flood tides. These maximum currents are toward the ocean, generating a set-down at the head. Likewise,
reduced near the bridge (at S4, with a maximum value of peaks in wind velocity that caused the greater set-up (from 22–
0.74 m/s). The velocity measurements from the mouth to the 25 June) had a persistent direction of around 2308 (SW winds),
bridge have a clear phase lag attributable to tidal propagation showing that these wind characteristics push water within the
in a frictional system. lagoon toward the head. The difference in these wind directions
is 1808. There is a period from 26–30 June, corresponding to
Numerical Experiments
diurnal breezes with a variable wind direction from 08 to 3308
Mainly because of the flat nature of the inland topography at
that exerts a small effect on water levels, of less than 2 cm at
the study area, winds are bound to have synoptic behavior (i.e.
the head. The effect of winds on surface elevation represents a
spatially homogeneous). Therefore, larger scale wind patterns
considerable fraction of the local tidal elevation at the head. In
are expected to be similar in direction and magnitude to local
this part of the lagoon, the measured diurnal tidal range was 11
winds. This characteristic generates uncertainties related to
cm on 24 June; therefore, according to this numerical
the role of local winds on the hydrodynamics of the lagoon and experiment, wind effects could be equivalent to 35.45% of the
its ability to generate a set-up/set-down in the water levels astronomical tide contribution to water elevation at the head.
since other processes might also play an important role. In this Nevertheless, including astronomical tides and the wind
sense, subtidal sea-level oscillations acting at the lagoon inlet forcing (MT-3) in the model has an unexpected effect on water
and related to remote atmospheric forcing could also be elevations, since maximum wind set-up at the head reduces to
responsible. Regardless of the exact mechanism operating on ~2.5 cm above the high tide compared to the corresponding
the continental shelf that causes the low-frequency fluctua- model without wind (MT-2). Inclusion of low-frequency
tions, its influence on the lagoon hydrodynamics is investigated
using a numerical modeling approach, as explained earlier.
The results are summarized in the following sections.
Figure 4. Wind rose for (a) the observation period May–July 2012 and b) the Figure 5. (a) Time series of measured water levels, (b) wind speed, and (c)
extended period November 2010–August 2011. wind direction for the observation period May–July 2012.

Table 6. Periods and amplitudes of astronomical tide constituents at Tide

Gauge (TG) station for the period March 2010–August 2011 (17 months).
Astronomical Tide
Constituent Period Amplitude (m)
SSA 1/2 year 0.0525
MSM 31.81 days 0.0266
MF 13.66 days 0.0242
O1 25.82 hours 0.1718
K1 23.93 hours 0.1905
N2 12.66 hours 0.0340 Figure 7. Frequency spectrum of sea level at Sea station for the observation
M2 12.42 hours 0.1223 period May–July 2012: (a) with one nonoverlapping interval, (b) with four
S2 12.00 hours 0.0421 nonoverlapping intervals.
oscillations further decreases wind set-up at the head to ~1.9 previously, winds do not considerably affect these circulation
cm, representing an increment of only 5% above the high tide patterns.
compared to the corresponding model without wind (MT-5
minus MT-4, for June 24th, 2010). This implies a reduction of Effects of Remote Forcing
~51% compared to the MT-1 test. The possible reasons for this The experiments that include measured sea levels at the
are addressed in the ‘‘Discussion.’’ As would be expected, the offshore open boundary (MT-4 and MT-5, without and with
addition of winds in MT-3 did not help to reproduce the low- wind, respectively) include the effects of remote forcing and,
frequency increase in water levels around 24 June at the head therefore, were able to simulate the low-frequency increase in
of the system (Figure 10); thus, it is clear that local winds are water level observed around 24 June along the lagoon (Figure
not the dominant forcing behind this important phenomenon. 13). The main difference compared to the two previous tests
Indeed, local winds play a small role in modifying the tidal (MT-2 and MT-3), which were forced with astronomical tides
signal and are only slightly appreciated at the head. (without and with wind, respectively), is that tides superim-
The effect of local winds on longitudinal velocities (tidal pose on the low-frequency fluctuation in water level caused by
currents) is practically negligible according to this 2D depth- the remote forcing. Similar to the model runs that include
averaged model (see Figure 11). Modeled velocities with or astronomical tides, the longitudinal velocities from the exper-
without wind have virtually the same behavior. Simulated iments that include the remote effects are not apparently
longitudinal velocities also reveal considerable damping at the affected by the low-frequency fluctuation.
head due to the closed nature of the system and friction. The Nevertheless, a careful analysis of volume fluxes and the
numerical model suggests the existence of remnant tidal cumulative water volumes (water that remains in or escapes
currents at the head of the system that are not observed in from the lagoon) reveals the importance of the low-frequency
the measurements. Flow vectors are plotted for the verified oscillations (Figure 14). Maximum instantaneous volume
model (MT-5) during ebb tide (Figure 12), showing several fluxes are reduced ~73%, from 508 m3/s at the mouth cross-
aspects of the depth-averaged tidal circulation. The model section to 137 m3/s at the bridge (Figure 14a). Volume fluxes
reproduces the maximum currents in the main channel, which are much greater at the mouth than at the bridge, not for its
collects water from the adjacent tidal flats (and the opposite greater cross-sectional area but because of an active tidal
occurs on the flood tide); currents intensify under the bridge
due to the reduction in the cross-section and diminish toward
the head, where horizontal currents are very small. Moreover,
the deviation of the currents by the embankment road can also
be appreciated in Figure 12b; the currents in the corners
between the lagoon border and the embankment road are
reduced, which explains the sedimentation observed in the
bathymetric map. The absence of tidal circulation at the head
causes practically no advective transport there. As mentioned
Figure 6. Time series of measured sea level at Sea station, calculated Figure 8. (a) Time series of sea level at Sea station and (b) measured
astronomical tide and residual (nonastronomical signal) from a harmonic longitudinal velocities for the observation period May–July 2012. Positive
analysis for the observation period May–July 2012. (Color for this figure is velocities are toward the head of the lagoon. (Color for this figure is available
available in the online version of this paper.) in the online version of this paper.)

Figure 9. (a) Simulated water levels for MT-1 (only wind forcing) compared
with (b) measured wind speed and (c) wind direction. Vertical dash lines are
used as a reference from selected wind speed peaks. Figure 11. Measured vs. simulated longitudinal depth-averaged velocities
for MT-4 and MT-5 (verified model): (a) at S1, (b) at S4, and (c) at S6. Positive
velocities are toward the head of the lagoon.
influence in the lower part of the lagoon where horizontal flows
are less affected by frictional effects. At the head of the system
the horizontal flows are reduced to near-zero values. Since without it, representing a reduction of ~14.4%. It should be
instantaneous volume fluxes are a function of longitudinal noted that the model without the embankment road was run
velocities, they are not apparently affected by the low- for the same bathymetry, which has been affected by the
frequency water-level fluctuation either. Nevertheless, if the sedimentation caused by the embankment road. Therefore,
cumulative water volumes that pass through the mouth and whereas the cumulated volumes are caused by the subtidal sea-
bridge cross-sections are calculated, the effects of the low- level fluctuation, their reduction from the mouth cross-section
frequency sea level increase are clearly evident (Figure 14b). to the bridge is attributed to the closed nature of the system at
Positive values in Figure 14b represent water that enters and the head, just as we have commented about for longitudinal
remains in the lagoon, and negative values are for water velocities and volume fluxes.
exiting the system. It is clear that the subtidal water-level Long-Term Analysis of the Remote Forcing
increase, generated by remote effects, causes a net entrance of Because the subtidal variability in water level seems to
water into the lagoon with a maximum cumulative volume of govern the accumulation (or loss) of water volume within the
12.27 3 106 m3, entering the system through the mouth cross- lagoon, it is important to know whether this phenomenon is a
section by 24 June. This then reduces to 6.242 3 106 m3 at the persistent occurrence in the region or whether it was only an
bridge, representing ~49% of the cumulated volume that isolated event observed during the field campaign of June 2012.
passed through the mouth. It is also of interest to understand the processes that can
It is worth mentioning that the reduction in the cumulative generate such behavior in the sea levels. For this purpose, a
volumes at the bridge cross-section is not attributed to the
embankment road that blocks the water flow in the middle zone
of the lagoon. This was confirmed by running a model without
this physical obstruction (implemented with the verified model
parameters but not presented in this paper), which shows that
cumulated volumes do not change. Only the longitudinal
velocities are locally reduced at S4, by flow continuity, from a
maximum of 0.423 m/s with the obstruction to 0.362 m/s
Figure 10. Measured vs. simulated water levels for MT-2 and MT-3, forced Figure 12. (a) Flow vectors for the verified model (MT-5) during the ebb tide;
with astronomical tides without and with local winds, respectively, (a) at S2, (b) zoom at the middle zone of the lagoon (bridge zone); and (c) zoom at the
(b) at S4, and (c) at S5. mouth zone.

Figure 13. Measured vs. simulated water levels for MT-4 and MT-5 (verified Figure 15. (a) Detrended measured sea level at TG station for the period
model) forced with measured sea level at Sea station without and with local March 2010–August 2011; (b) high-pass and low-pass Fourier filters with a
winds, respectively, (a) at S2, (b) at S4, and (c) at S5. 7-day cutoff period.
sea-level record of ~17 months (March 2010 to August 2011) This means that a wind blowing northward (southward)
measured at the Celestun TG was used. Figure 15a shows the consistently coincides with sea-surface elevations (depres-
entire sea-level signal, which was filtered (Fourier) using a 7- sions), reinforcing the idea that a synoptic large scale wind
day cutoff period to obtain the high-pass and low-pass could act on the coastal ocean through Coriolis leading to
variability (Figure 15b). The low-pass data reveals that the water-level variations at the coast (Ekman transport).
subtidal sea-level oscillations are very common features of the The results of the longer term simulation (MT-6), spanning
Yucatan coast occurring during the entire year but are of larger two months, are presented in Figure 18 to corroborate the
amplitude during autumn and winter. Harmonic analysis of results of the previous subsection, but we are now testing the
this signal (Table 6) shows that the long-term tidal constituents effects of a low-frequency sea-level depression. The results
(six monthly, monthly, and lunar fortnightly) are not negligible consist of simulated water levels and cumulative water
and can substantially contribute to the lower frequencies. A volumes through the cross-sections (mouth and bridge). As
frequency spectrum of the residual (nonastronomical) signal is would be expected, the low-frequency water levels profoundly
presented in Figure 16 and shows significant low-frequency influence the water exchange between the coastal ocean and
oscillations (peaks) at 19.5, 30.34, and 68.26 days. This result the lagoon. On 14 January 2011, the subtidal depression causes
confirms that low-frequency oscillations of astronomical and 11.38 3 106 m3 of water to exit the system through the mouth
nonastronomical origin can contribute substantially to sea- cross-section. Figure 18b also demonstrates that tides are very
level variability and that the 21-day oscillation observed in efficient at transporting water through the cross sections, as
June 2012 falls within these spectral peaks, being of nonastro- the diurnal variability is quite large. Nevertheless, the
nomical origin. The effects of subtidal sea-level fluctuations on presence of the low frequency will make the difference between
the lagoon hydrodynamics are bound to be very similar balance, preferential export, or import of water through the
regardless of the origin of the low-frequency oscillation system.
(astronomical, Ekman transport, coastally trapped waves,
etc.). In spite of this, a contribution to the understanding of DISCUSSION
the mechanisms responsible for these oscillations is important. The results from measured data and modeling experiments
Hence, a cross-spectrum was performed between the residual show that there is an important difference in the magnitudes of
(nonastronomical) sea level at TG and the along-shore the water-level vertical displacements caused by remote
component of the wind, where positive values are northward forcing vs. local winds. At the head of the lagoon, a comparison
winds. A positive and coherent relationship is found between between the results of MT-2 (astronomical tide without winds)
the residual sea-level and the along-shore winds (Figure 17). and MT-4 (measured sea level without winds) shows that the
remote effects account for an elevation of ~26.6 cm (320% of the
measured diurnal tidal elevation) for the event of 24 June. On
the other hand, the model forced solely with realistic local
Figure 14. (a) Simulated volume fluxes and (b) cumulated water volumes Figure 16. Frequency spectrum of the residual tide at TG station for the
through cross-sections for the verified model (MT-5). Positive values are extended period March 2010–August 2011 with two nonoverlapping
toward the head of the lagoon. intervals.

Figure 18. (a) Simulated water levels and (b) cumulated water volumes
through cross-sections for MT-6. Positive volumes are toward the head of the
Figure 17. Coherence analysis between along-shore winds and residual tide lagoon.
for the extended period (TG station) from March 2010–August 2011: (a)
cospectrum with eight nonoverlapping intervals, (b) square coherence.
superelevation observed in the lagoon on 24 June 2012 cannot
be attributable to local wind effects and must be related to a
winds (MT-1) revealed a maximum set-up at the head of only remote influence acting on the inlet. This is an interesting
3.9 cm during the same period. MT-1 suggests that although finding because it could otherwise be interpreted as an effect of
the effects of local winds on water-level variability are small, local winds aligned to the main axis of this elongated lagoon.
they can represent an important contribution (35.45%) of the This is also supported by the fact that this superelevation was
measured diurnal tidal range (0.11 m) at the head. Neverthe- observed in the open ocean, 5 km offshore from the inlet, away
less, it is interesting to note that wind influence reduces to ~1.9 from the influence of the lagoon. Wong and Moses-Hall (1998)
cm when including the diurnal tides and low-frequency found that given the large difference in size between an estuary
fluctuation (MT-5 minus MT-4). This could suggest that (or lagoon) and the coastal ocean, it is reasonable to assume
diurnal local winds do not effectively couple with diurnal tides that the sea level will not be affected by the processes operating
to cause an important set-up at the head; on the contrary, tides in the estuary, with sea-level variations being produced by
seem to counteract the wind effects. A possible explanation is winds over the continental shelf adjacent to the estuary (or
that local winds interact with the diurnal tides and the low- lagoon) (i.e. remote winds). Subsequently, the impingement of
frequency oscillation, as suggested by Wong and Trowbridge the sea level at the mouth transmits these effects into the
(1990), through nonlinear interaction between the tidal wave system. Using a simple barotropic model, Garvine (1985) found
and the currents in the turbulent boundary layer, enhancing that the wavelength of the coastally forced subtidal fluctua-
frictional attenuation. The local set-up/set-down observed at tions is much longer than the length of most estuaries;
MT-1 at the head with a surface slope along the lagoon, therefore, the coastally forced sea levels within the estuary
however, is an effect that has also been documented by other closely follow those at the mouth without any phase lag. This
authors (e.g., Garvine, 1985; Mied et al., 2010; Möller et al., means that the low-frequency fluctuation at the inlet transmits
2001; Sankaranarayanan and Fringer, 2013; Santoro, Fossati, to the whole lagoon with very little attenuation. At Celestun,
and Piedra-Cueva, 2013; Wong and Moses-Hall, 1998). For the an attenuation of ~17% from S5 to S2 for the low-frequency
elongated lagoon of Puttalam, Sri Lanka, Wijeratne and signal (21.33 days) was observed. In comparison, the tidal wave
Rydberg (2007) found a wind set-up/set-down of 8 cm at the attenuation at a diurnal frequency is ~85% from the mouth
head for moderate constant winds (which might overestimate (S6) to the head (S2) of the lagoon. These values emphasize the
effects) of 4 m/s. Similarly, model simulations carried out by large frictional characteristics of the lagoon because of its
Sankaranarayanan and Fringer (2013) in San Francisco Bay shallowness and extensive underwater vegetation cover.
show that wind set-up is around 1–3 cm. On the other hand, the It is also interesting to understand what processes are
low-frequency water-level displacement caused by the remote driving the remote effect at the site. The analyses of a longer
forcing at Celestun is similar to that shown by Wong and Valle- term time series (17 months) show that several processes
Levinson (2002) of up to 0.4 m above msl. Although the effect of contribute to the sea-level variability at subtidal frequencies at
remote winds seems to have a much greater effect than local Celestun. Harmonic analysis clearly shows that the lunar
winds on the water levels inside lagoons in several studies, fortnightly (MF), monthly (MSM), and six-monthly (SSA)
Garvine (1985) argues that the effects of the remote and local components are important. Nevertheless, periods closer to 20
winds operate either in concert or in opposition, depending on days (as observed during the field campaign) are clearly related
hemisphere and orientation of the estuary axis relative to the to the residual (nonastronomical) component. A cross-spectrum
coast and the wind. In this regard, Cho et al. (2012) found that between the residual sea-surface elevation and the along-shore
hurricane Floyd was followed by down-Bay winds in the component of the wind suggests that the low-frequency
Chesapeake Bay that canceled the initial set-up (originated oscillations at such frequencies are associated with an along-
by the remote winds) and caused a set-down from the upper shelf process caused by Ekman transport where winds from the
estuary. SW can pile up water toward the coast by a cross-shelf forcing
In our case, considering the numerical tests presented in the associated with Coriolis effects. Note that the opposite would
results and the known behavior of subtidal water levels in apply for the subtidal water-level depressions, a process that is
coastal water bodies, it is quite clear that the low-frequency actually more likely to occur given the persistence of NE

alongshore winds. This process has been observed in several export of materials and properties from the lagoon to the
previous studies (e.g., Möller et al., 2001; Smith 1977; Wong coastal ocean. Because the remote forcing is so important for
and Moses-Hall, 1998; Wong and Valle-Levinson, 2002). the lagoon hydrodynamics, numerical models forced solely with
Regardless of the nature of the remote forcing, this astronomical tides (with or without local winds) would deviate
phenomenon will profoundly influence the water levels inside from reality, having very limited capabilities to account for
this coastal lagoon. But what about the other hydrodynamic flooding scenarios.
properties? Sankaranarayanan and Fringer (2013) found that
at San Francisco Bay, the remote effects are more effective CONCLUSIONS
than direct (local) wind stress at generating low-frequency The influence of remote and local winds, as well as
fluctuations in water levels inside the lagoon but did not affect astronomical tides on the water levels and longitudinal
water-surface slope. Furthermore, the effects of local winds on velocities at tidal and subtidal frequencies, were investigated
water-surface slope were essential enough to explain the in the coastal lagoon of Celestun, Mexico. This was achieved by
behavior of subtidal circulation patterns. At this site, subtidal means of field observations and the implementation of a
variability in lagoon velocity components was more related to barotropic 2D numerical model. The calibration and verifica-
local winds, despite the fact that remote forcing had a greater tion of the numerical model resulted in good to excellent E,
influence on water levels. For the case of Celestun lagoon, RMAE, and REV values; therefore, the model is considered
neither local winds nor remote effects had a noticeable reliable for reproducing the hydrodynamic behavior of the
influence on longitudinal velocities, but it was clear that the lagoon.
subtidal fluctuations had a profound effect on the cumulative The tidal signal propagates from the inlet toward the lagoon
water volumes of the lagoon. For example, the subtidal head, being attenuated ~85% with a lag time of ~7.75 hours.
superelevations observed for MT-5 presented an accumulated Tidal currents also reveal damping and a lag time from the
water volume of 12.27 3 106 m3, which was imported from the inlet to the head, being very small at the head due to the closed
coastal ocean, whereas the maximum low-frequency decreases nature of the system. Field measurements and data analysis
in sea level of MT-6 exported 11.38 3 106 m3 to the coastal show the presence of a low-frequency fluctuation in the water
ocean. Wong (1986, 1987) suggested that volume exchanges at level, which is identified to be of remote origin. This long wave
tidal frequencies could not be as effective as those at subtidal is readily transmitted toward the lagoon head with little
frequencies because tidal energy is more easily dissipated than attenuation (~17%) and is only reproduced in the numerical
low-frequency sea-level fluctuations. Our results (Figure 18) model when forced at the offshore open boundary with sea-
show that the variability in cumulative volumes at tidal level measurements. Further analysis of longer term sea-level
frequencies is very high and in the same order of magnitude data (17 months) shows that this type of oscillation is very
as the low-frequency variability. This implies that a rough common at the site and can be linked to astronomical
balance exists between the water volumes entering and exiting components (SSA, MSM, MF) and to the residual (nonastro-
the lagoon if sea level is not influenced by low-frequency nomical) elevations. The latter have characteristic periods of
oscillations, however, as soon as a depression occurs, the ~20, 30, and 68 days and are positively correlated with the
volume of water is preferentially exported. Similarly, when alongshore wind components, which suggests they must be
low-frequency sea-level increases occur, seawater is preferen- linked to Ekman transport at the shelf. Local winds can also
tially imported into the lagoon. cause subtidal water-level fluctuations, depending on wind
This has important implications for the site. For example, intensity, duration, and persistence in the along-axis direction.
several extraordinary flooding events have occurred along the Numerical experiments suggest that local SW winds (i.e.
margins of Celestun lagoon, which have socio-economic toward the head) acting in the absence of other forcings can
consequences, particularly for the ecotourism pier facilities contribute to the remote effect, generating a set-up at the head
located on the lagoon. Results from this study suggest that of the system of up to 3.9 cm. Similarly, NE winds (i.e. toward
these extreme events could not have been caused by local winds the ocean) cause a set-down of about 5.9 cm. However, when
since these events generate a relatively small set-up at the winds and astronomical tides coexist, as is usually the case, the
head (of the order of a few cm); rather, the floods could have effects of wind on the surface elevations seem to be attenuated
been caused by a low-frequency sea-level increase generated by (51%). Thus, local winds play a small role in modifying the tidal
remote winds or by any other process that affects sea levels in signal and are only slightly distinguishable at the head of the
the region. For example, based on observations of sea-surface lagoon at low frequencies.
elevation at several ports along the Gulf of Mexico, it is clear It is interesting to note that sea-level low-frequency
that an annual fluctuation that increases sea level up to 15–20 fluctuations do not affect tidal currents and volume fluxes
cm above msl occurs from October to November. This sea- because the main variability is of high frequency (diurnal);
surface anomaly has been attributed to the along-coast wind- however, it causes a subtidal fluctuation on the cumulative
stress component and to the low-frequency variability of volumes through the cross-sections with a maximum cumula-
atmospheric pressure (Zavala-Hidalgo, Morey, and O’Brien, tive water volume of 12.27 3 106 m3, which entered the system
2003). Mateos-Jasso (2009) also suggested that it could be through the mouth by 24 June 2012. Accordingly, sea-level
related to coastal waves generated as a consequence of changes depressions, caused by a NE wind, would contribute consider-
in wind direction. Alternatively, when the opposite effect ably to water export from the lagoon to the coastal sea, as
occurs (sea-level decreases), an excess outflow should exist, observed in the long-term numerical experiment. This stresses
with important implications for the water renewal and the the importance of low-frequency fluctuations in water exchang-

es between the ocean and the lagoon, with important Herrera-Silveira, J.A., 1994. Nutrients from underground water
implications for the exchange processes. Thus, to account for discharges in a coastal lagoon (Celestun, Yucatan, Mexico).
Verhandlungen des Internationalen Verein Limnologie, 25(1),
any subtidal fluctuation in the water levels, it is important to 1398–1401.
include remote forcing in the models by prescribing measured Herrera-Silveira, J.A., 1996. Salinity and nutrients in a tropical
water levels at the sea boundary and not just astronomical coastal lagoon with groundwater discharges to the Gulf of Mexico.
tides, especially for flooding simulations. Hydrobiologı́a, 321(2), 165–176.
Herrera-Silveira, J.A.; Troccoli-Ghinaglia, L.; Ramirez-Ramirez, J.,
and Zaldivar-Jimenez, A., 1999. Yucatan groundwater-influenced
ACKNOWLEDGMENTS systems: Laguna de Celestun, Yucatan. In: Smith, S.V.; Crossland,
This work was partly supported by the National Council of J.I.M., and Crossland, C.J., (eds.), Mexican and Central American
Science and Technology (Conacyt, Mexico) and the Yucatan Coastal Lagoon Systems: Carbon, Nitrogen and Phosphorus Fluxes
(Regional Workshop III). Texel, The Netherlands: LOICZ Reports
government through the projects YUC-2006-C05-66254 (FO-
& Studies, pp. 19–26.
MIX) and 101720 (‘‘Geoquı́mica de metales en lagunas Hu, J.; Gao, J., and Wang, X., 2009. Multifractal analysis of sunspot time
costeras: factores directrices y modelado’’). RC-S had a series: The effects of the 11-year cycle and Fourier truncation. Journal
doctoral grant provided by Conacyt. The authors gratefully of Statistical Mechanics: Theory and Experiment, 2009(2), 1–20.
Iglesias, G. and Carballo, R., 2010. Effects of high winds on the
acknowledge the Mexican Navy (SEMAR) for providing SPOT
circulation of the using a mixed open boundary condition: The Rı́a de
satellite imagery, as well as Dr. David Valdés-Lozano Muros, Spain. Environmental Modelling & Software, 25(4), 455–466.
(Cinvestav) and Biol. David Alonso-Parra (Dumac) and Janzen, C. and Wong, K-C., 1998. On the low-frequency transport
CONAGUA (the Mexican water management agency) for processes in a shallow coastal lagoon. Estuaries, 21(4), 754–766.
providing the meteorological data. The authors wish to Jia, P. and Li, M., 2012. Dynamics of wind-driven circulation in a
shallow lagoon with strong horizontal density gradient. Journal of
express their sincere thanks to Dr. Paulo Salles-Afonso de Geophysical Research, 117(C5), 1–14.
Almeida (UNAM) for providing TG records from Celestun Kankara, R.S.; Mohan, R., and Venkatachalapathy, R., 2013.
pier, Dra. Cecilia Enriquez-Ortiz (UNAM) for her support in Hydrodynamic modelling of Chennai coast from a coastal zone
the field campaigns, on the data processing, and for the management perspective. Journal of Coastal Research, 29(2), 347–
357.
constructive comments on this paper, and to Biol. Emmanuel Lankford, R.R., 1977. Coastal lagoons of Mexico: Their origin and
Uc-Sánchez (Cinvestav) for his support in the field and data classification. In: Wiley, M.L. (ed.), Estuarine Processes. New York:
processing. The authors would like to thank the anonymous Academic, pp. 182–215.
reviewers for their very helpful comments that improved this Lesser, G.R.; Roelvink, J.A.; van Kester, J.A.T.M., and Stelling, G.S.,
2004. Development and validation of a three-dimensional morpho-
contribution substantially. logical model. Coastal Engineering, 51(8–9), 883–915.
Luijendijk, A., 2001. Validation, Calibration and Evaluation of a
LITERATURE CITED Delft3D-FLOW Model with Ferry Measurements. Delft, The
Barrios-Espino, G.R., 1988. Aspects of the Ecology of the Caribbean Netherlands: Technical University of Delft, Master’s thesis, 92p.
Flamingo (Phoenicopterus ruber ruber) in Yucatan, Mexico. Malhadas, M.S.; Neves, R.J.; Leitão, P.C., and Silva, A., 2010.
Influence of tide and waves on water renewal in Óbidos Lagoon,
Auburn, Alabama: Auburn University, Master’s thesis, 65p.
Portugal. Ocean Dynamics, 60(1), 41–55.
Cho, K-H.; Wang, H.V.; Shen, J.; Valle-Levinson, A., and Teng, Y., 2012.
Mateos-Jasso, A., 2009. Estudio sobre la Estacionalidad de la
A modeling study on the response of Chesapeake Bay to hurricane
Surgencia de Yucatán. México, D.F.: Universidad Nacional
events of Floyd and Isabel. Ocean Modelling, 49–50(1), 22–46.
Autónoma de México, Master’s thesis, 102p.
Chow, V.T., 1959. Open Channel Hydraulics. New York: McGraw-
Mied, R.P.; Schulz, W.J.; Handler, R.A.; Snow, C.M.; Fusina, R.A.,
Hill, 680p.
and Porter, J.H., 2010. Remote and local forcing of a coastal lagoon:
De Marchis, M.; Ciraolo, G.; Nasello, C., and Napoli, E., 2012. Wind-
The Virginia Coast Reserve. Continental Shelf Research, 30(20),
and tide-induced currents in the Stagnone lagoon (Sicily). Envi-
2057–2066.
ronmental Fluid Mechanics, 12(1), 81–100.
Möller, O.O.; Castaing, P.; Salomon, J-C., and Lazure, P., 2001. The
De Marchis, M.; Freni, G., and Napoli, E., 2014. Three-dimensional Influence of local and non-local forcing effects on the subtidal
numerical simulations on wind- and tide-induced currents: The case circulation of Patos lagoon. Estuaries, 24(2), 297–311.
of Augusta Harbour (Italy). Computers & Geosciences, 72(1), 65–75. Nash, J.E. and Sutcliffe, J.V., 1970. River flow forecasting through
Deltares, 2011. Delft3D-FLOW. Simulation of Multi-Dimensional conceptual models, part I—A discussion of principles. Journal of
Hydrodynamic Flows and Transport Phenomena, Including Sedi- Hydrology, 10(3), 282–290.
ments. User Manual,. Hydro-Morphodynamics. Delft, The Nether- Pawlowicz, R.; Beardsley, B., and Lentz, S., 2002. Classical tidal
lands: Deltares, Version 3.15, 674p. harmonic analysis including error estimates in MATLAB using
Erskine, A.D., 1991. The effect of tidal fluctuation on a coastal aquifer T_TIDE. Computers and Geosciences, 28(8), 929–937.
in the UK. Ground Water, 29(4), 556–562. Pech, D.; Ardisson, P.L., and Hernández-Guevara, N.A., 2007.
Fernandes, E.H.L.; Mariño-Tapia, I.; Dyer, K.R., and Möller, O.O., Benthic community response to habitat variation: A case of study
2004. The attenuation of tidal and subtidal oscillations in the Patos from a natural protected area, the Celestun coastal lagoon.
Lagoon estuary. Ocean Dynamics, 54(3–4), 348–359. Continental Shelf Research, 27(20), 2523–2533.
Fischer, H.B.; List, E.G.; Koh, R.C.Y.; Imberger, J., and Brooks, N.H., Perry, E.; Velazquez-Oliman, G., and Marin, L., 2002. The hydro-
1979. Mixing in Inland and Coastal Waters. New York: Academic geochemistry of the karst aquifer system of the northern Yucatan
Press, 483p. Peninsula, Mexico. International Geology Review, 44(3), 191–221.
Garvine, R.W., 1985. A simple model of estuarine subtidal fluctua- Ramsar, 2015. Ramsar Sites Information Service. https://rsis.ramsar.
tions forced by local and remote wind stress. Journal of org/.
Geophysical Research, 90(C6), 11945–11948. Roelvink, J.A. and Walstra, D.J.R., 2004. Keeping it simple by using
Gerritsen, H.; de Goede, E.D.; Platzek, F.W.; Genseberger, M.; van complex models. Proceedings of the 6th International Conference on
Kester, J.A.T.M., and Uittenbogaard, R.E., 2007. Validation Hydro-Science and Engineering. Advances in Hydro-Science and
Document Delft3D-FLOW; a Software System for 3D Flow Engineering Vol. VI (Brisbane, Australia), pp. 1–11.
Simulations. The Netherlands: Delft Hydraulics, Report X0356, Sánchez-Arguelles, R.D., 1994. Comparación Estructural de la
M3470, 266p. Comunidad de Manglar en Dos Sistemas Lagunares Costeros del

Estado de Yucatán, M éxico. Yucatán, M éxico: Universidad COAST3D Data at Teignmouth 1999. Oxfordshire, U.K.: HR
Autónoma de Yucatán, Bachelor’s thesis, 67p. Wallingford, TR121-EC MAST Project No. MAS3-CT97-0086, 4p.
Sankaranarayanan, S. and Fringer, O.B., 2013. Dynamics of Walstra, D.J.R.; van Rijn, L.C.; Boers, M., and Roelvink, J.A., 2003.
barotropic low-frequency fluctuations in San Francisco Bay during Offshore sand pits: Verification and application of a hydrodynamic
upwelling. Continental Shelf Research, 65(1), 81–96. and morphodynamic model. Proceedings of the 5th International
Santoro, P.E.; Fossati, M., and Piedra-Cueva, I., 2013. Study of the Conference on Coastal Sediments (Clearwater, USA), pp. 1–14.
meteorological tide in the Rı́o de la Plata. Continental Shelf Wang, D.P. and Elliott, A.J., 1978. Non-tidal variability in the
Research, 60(1), 51–63. Chesapeake Bay and Potomac River: Evidence for non-local
SARH (Secretarı́a de Agricultura y Recursos Hidráulicos), 1989. forcing. Journal of Physical Oceanography, 8(2), 225–232.
Datos Climatológicos del Norte de Yucatán. México, D.F.: Secre-
Waterhouse, A.F.; Tutak, B.; Valle-Levinson, A., and Sheng, Y.P.,
tarı́a de Agricultura y Recursos Hidráulicos, Reporte Anual, 25p.
2013. Influence of two tropical storms on the residual flow in a
Shetye, S.R. and Vijith, V., 2013. Sub-tidal water-level oscillations in
subtropical tidal inlet. Estuaries and Coasts, 36(5), 1037–1053.
the Mandovi estuary, west coast of India. Estuarine, Coastal and
Shelf Science, 134(1), 1–10. Wijeratne, E.M.S. and Rydberg, L., 2007. Modelling and observations
Smith, N.P., 1977. Meteorological and tidal exchanges between of tidal wave propagation, circulation and residence times in
Corpus Christi Bay, Texas and the northwestern Gulf of Mexico. Puttalam Lagoon, Sri Lanka. Estuarine, Coastal and Shelf Science,
Estuarine and Coastal Marine Science, 5(4), 511–520. 74(4), 697–708.
Smith, N.P., 1978. Long-period, estuarine-shelf exchanges in re- Wong, K-C., 1986. Sea-level fluctuations in a coastal lagoon.
sponse to meteorological forcing. In: Jacques, C.J.N. (ed.), Proceed- Estuarine, Coastal and Shelf Science, 22(6), 739–752.
ings of the 9th International Liege Colloquium on Ocean Wong, K-C., 1987. Tidal and subtidal variability in Delaware’s inland
Hydrodynamics, Hydrodynamics of Estuaries and Fjords. Amster- bays. Journal of Physical Oceanography, 17(3), 413–422.
dam, The Netherlands, pp. 147–159. Wong, K-C. and Moses-Hall, J.E., 1998. On the relative importance of
Sousa, M.C. and Dias, J.M., 2007. Hydrodynamic model calibration the remote and local wind effects to the subtidal variability in a
for a mesotidal lagoon: The case of Ria de Aveiro (Portugal). coastal plain estuary. Journal of Geophysical Research, 103(C9),
Proceedings of the 9th International Coastal Symposium (Gold 18393–18404.
Coast, Australia). Journal of Coastal Research, Special Issue No. Wong, K-C. and Trowbridge, J., 1990. Some observational evidence on
50, pp. 1075–1080. the effect of atmospheric forcing on tidal variability in the Upper
Van Rijn, L.C.; Walstra, D.J.R.; Grasmeijer, B.T.; Sutherland, J.; Pan, Delaware Bay. Journal of Geophysical Research, 95(C9), 16229–
S., and Sierra, J.P., 2002. Simulation of nearshore hydrodynamics
16240.
and morphodynamics on the time scale of storms and seasons using
Wong, K-C. and Valle-Levinson, A., 2002. On the relative importance
process based profile models. In: Van Rijn, L.C.; Ruessink, B.G.,
of the remote and local wind effects on the subtidal exchange at the
and Mulder, J.P.M., (eds.), The Behaviour of a Straight Sandy
Coast on the Time Scale of Storms and Seasons: Process Knowledge entrance to the Chesapeake Bay. Journal of Marine Research,
and Guidelines for Coastal Management. Amsterdam, The Nether- 60(3), 477–498.
lands: Coast3D-Egmond, pp. 816–831. Wong, K-C. and Wilson, R., 1984. Observations of low frequency
Walstra, D.J.R.; van Rijn, L.C., and van Helvert, M.A.G., 2002. variability in Great South Bay and relations to atmospheric
Morphology of Pits, Channels and Trenches. Part II: Model forcing. Journal of Physical Oceanography, 14(12), 1893–1900.
Verification of Delft3D with PUTMOR Dataset. The Netherlands: Zavala-Hidalgo, J.; Morey, S.L., and O’Brien, J.J., 2003. Seasonal
Delft Hydraulics, Report Z3223, 149p. circulation on the western shelf of the Gulf of Mexico using a high-
Walstra, D.J.R.; van Rijn, L.C.; Blogg, H., and van Ormondt, M., resolution numerical model. Journal of Geophysical Research,
2001. Evaluation of a Hydrodynamic Area Model Based on the 108(C12), 19-1–19-19.
A RESUMEN A
El principal objetivo de este estudio es investigar los efectos de los vientos locales y las variaciones del nivel del mar en las frecuencias mareales y
submareales de los niveles de agua, velocidades promediadas en la profundidad e intercambios de agua con el mar en una laguna costera alargada y somera,
con base en mediciones de campo y un modelo numérico barotrópico en 2D. Las mareas son principalmente diurnas y se atenúan ~85% desde la boca hasta la
cabeza de la laguna indicando una propagación barotrópica dominada por fricción. Por otro lado, las fluctuaciones de baja frecuencia en los niveles de agua se
propagan con poca atenuación (~17%) dominando la variabilidad del nivel del agua en la cabeza. Es importante elucidar si esta variabilidad de baja
frecuencia está relacionada con los vientos locales o se debe a los efectos remotos. Los resultados de modelación muestran que los vientos locales pueden
incrementar los niveles de agua hasta ~1.9 cm en la cabeza de la laguna (5% sobre el nivel de marea), mientras que la contribución de la fluctuación de baja
frecuencia fue de hasta ~26.6 cm (320% sobre las mareas astronómicas), controlando los niveles de agua dentro de la laguna a frecuencias submareales. Por
otro lado, las brisas marinas de frecuencia diurna juegan un rol menor en los niveles de agua dentro de la laguna y no refuerzan a las mareas diurnas para
causar un incremento importante del nivel del agua. Los volúmenes acumulados a través de secciones transversales están altamente correlacionados con el
nivel del mar y muestran que las fluctuaciones de baja frecuencia del nivel del mar son especialmente eficientes para acumular o exportar agua de la laguna
durante perı́odos de ascenso y descenso de esta variable, respectivamente. Esto implica que los intercambios de agua entre la laguna y el océano costero están
dominados no solo por las mareas diurnas sino principalmente por las fluctuaciones de baja frecuencia del nivel del mar, cuyos orı́genes son astronómicos y
meteorológicos (i.e. forzamiento del viento a través del transporte de Ekman).

Jcoastres D 14 00146.1

Uploaded by

Copyright:

Available Formats

Jcoastres D 14 00146.1

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Jcoastres D 14 00146.1

Uploaded by

Copyright:

Available Formats

Influence of the Remote Forcing and Local Winds on the Barotropic

Hydrodynamics of an Elongated Coastal Lagoon

Influence of the Remote Forcing and Local Winds on the

INTRODUCTION Although tides can be important drivers for the dynamics of

important gradients along the water surface, affecting currents

Journal of Coastal Research, Vol. 32, No. 1, 2016

Figure 3. Temperature and salinity profiles in the lagoon.

Journal of Coastal Research, Vol. 32, No. 1, 2016

Journal of Coastal Research, Vol. 32, No. 1, 2016

Manning Roughness Coefficients (n)

Journal of Coastal Research, Vol. 32, No. 1, 2016

Table 2. Numerical experiments summary.

Open Boundary Conditions

Journal of Coastal Research, Vol. 32, No. 1, 2016

Water Levels Velocities

Journal of Coastal Research, Vol. 32, No. 1, 2016

Measured Water Levels Simulated Water Levels

Journal of Coastal Research, Vol. 32, No. 1, 2016

Table 6. Periods and amplitudes of astronomical tide constituents at Tide

Journal of Coastal Research, Vol. 32, No. 1, 2016

Journal of Coastal Research, Vol. 32, No. 1, 2016

Journal of Coastal Research, Vol. 32, No. 1, 2016

Journal of Coastal Research, Vol. 32, No. 1, 2016

Journal of Coastal Research, Vol. 32, No. 1, 2016

Journal of Coastal Research, Vol. 32, No. 1, 2016

Journal of Coastal Research, Vol. 32, No. 1, 2016

You might also like