Marine Pollution Bulletin 105 (2016) 277–285
Contents lists available at ScienceDirect
Marine Pollution Bulletin journal homepage: www.elsevier.com/locate/marpolbul
Research paper
The density-driven circulation of the coastal hypersaline system of the Great Barrier Reef, Australia Gerry G. Salamena a,⁎, Flávio Martins b, Peter V. Ridd c a b c
College of Marine and Environmental Science, James Cook University, Townsville, Queensland 4811, Australia CIMA-EST/UAlg., Campus da Penha, P8000-117, Faro, Portugal Marine Geophysics Laboratory, College of Science Technology and Engineering, James Cook University, Townsville, Queensland 4811, Australia
a r t i c l e
i n f o
Article history: Received 16 September 2015 Received in revised form 30 January 2016 Accepted 4 February 2016 Available online 12 February 2016 Keywords: Density-driven circulation Hypersaline waters Great Barrier Reef Flushing time
a b s t r a c t The coastal hypersaline system of the Great Barrier Reef (GBR) in the dry season, was investigated for the first time using a 3D baroclinic model. In the shallow coastal embayments, salinity increases to c.a. 1‰ above typical offshore salinity (~35.4‰). This salinity increase is due to high evaporation rates and negligible freshwater input. The hypersalinity drifts longshore north-westward due to south-easterly trade winds and may eventually pass capes or headlands, e.g. Cape Cleveland, where the water is considerably deeper (c.a. 15 m). Here, a pronounced thermohaline circulation is predicted to occur which flushes the hypersalinity offshore at velocities of up to 0.08 m/s. Flushing time of the coastal embayments is around 2–3 weeks. During the dry season early summer, the thermohaline circulation reduces and therefore, flushing times are predicted to be slight longer due to the reduced onshore-offshore density gradient compared to that in the dry season winter period. © 2016 Elsevier Ltd. All rights reserved.
1. Introduction The existence of hypersaline waters in coastal zones of continental shelves is caused by the excess of evaporation over precipitation and river runoff (de Silva Samarasinghe and Lennon, 1987; Gräwe et al., 2009; Heggie and Skyring, 1999; Lavı́n et al., 1998; Wolanski, 1986). Hypersaline systems in continental shelves have been studied in numerous locations particularly around the dry continent of Australia, for example, in northern Australia (Wolanski, 1986), gulfs in the Southern Australia (de Silva Samarasinghe, 1989; de Silva Samarasinghe and Lennon, 1987; Nunes and Lennon, 1986, 1987), Hervey Bay, Australia (Gräwe et al., 2009) and coastal zones of Great Barrier Reef (GBR), Australia (Andutta et al., 2011; Wang et al., 2007; Wolanski, 1981). The extent of the hypersalinity depends upon the freshwater balance of the region which is highly associated with seasonal characteristics e.g. during the dry season summer (e.g. in Gulf of California, Lavı́n et al. (1998)) and the dry winter in the GBR (Walker, 1981b). One important hydrodynamical effect of the hypersaline waters is to produce a thermohaline circulation by which the saltier water masses will sink and be flushed out seaward along the sea-bed (Fig. 1) (Gräwe et al., 2009; Heggie and Skyring, 1999; Lennon et al., 1987b; Wolanski, 1986). As a result, this thermohaline circulation is an important oceanographic aspect in the hypersaline coastal waters of the continental shelf. ⁎ Corresponding author at: LIPI's Centre for Deep Sea Research, Indonesian Institute of Sciences (LIPI), Jl. Y. Syaranamual, Guru-Guru, Poka, Kota Ambon, Provinsi Maluku 97233, Indonesia. E-mail address: [email protected] (G.G. Salamena).
http://dx.doi.org/10.1016/j.marpolbul.2016.02.015 0025-326X/© 2016 Elsevier Ltd. All rights reserved.
The degree of hypersalinity on the continental shelf is not only influenced by the freshwater balance, but is also affected by the exchange of the coastal hypersaline waters with oceanic salinity from offshore (Wang et al., 2007). This exchange transport process is likely affected by turbulent diffusion, thermohaline circulation and large scale advection (Wang et al., 2007). Due to this exchange transport role, some authors have conducted studies connecting this transport process with the flushing time of the coastal hypersaline waters (de Silva Samarasinghe and Lennon, 1987; Hancock et al., 2006; Heggie and Skyring, 1999; Largier et al., 1997; Wang et al., 2007). The hypersaline system of the GBR shelf is different to most other hypersaline environments due to its aspect ratio. It is a continental shelf system 2000 km in the long-shelf direction and between 50 (further North) and 100 (further South) km across shelf. This contrasts with most other reported hypersaline systems which are bays, gulfs, or inverse estuaries and are smaller in the long-shelf direction than across shelf direction. Due to this morphological difference, along shore currents become important for the GBR (Andutta et al., 2011). In contrast, for waters in narrow bays, it is the cross shelf tidal currents which predominate e.g. Gulf of St. Vincent (de Silva Samarasinghe and Lennon, 1987), Gulf of California (Lavı́n et al., 1998), Shark Bay (Nahas et al., 2005) and San Diego Bay (Largier et al., 1997). Thus, along-shore currents of the GBR enables the hypersaline waters in the GBR to be transported alongshore from one coastal embayment to another (Andutta et al., 2011). There has been considerable works on residence or flushing times of the GBR waters due to the potential influence of residence time on pollutant build-up in the GBR lagoon (Andutta et al., 2013; Choukroun
278
G.G. Salamena et al. / Marine Pollution Bulletin 105 (2016) 277–285
Fig. 1. Schematic diagram of evaporation-driven circulation. The light grey indicates plume of hypersaline water. This figure is drawn based on Fig. 2d of Wang et al. (2007).
et al., 2010; Hancock et al., 2006; Mao and Ridd, 2015; Wang et al., 2007). For the inshore hypersaline system in the GBR, it is probable that a combination of longshore currents and cross-shelf transport including via the thermohaline circulation can significantly affect the exchange process between inshore and offshore waters. This contrasts with the situation in narrow bays and estuaries where the effects of longshore currents are negligible (Largier et al., 1997) and hence, the narrow waters are likely to have longer residence time (Kämpf et al., 2010). Recent studies investigating cross-shelf transport in the GBR's coastal hypersaline system, which used 1D-models of the diffusion of salt, indicate short residence times (a few weeks) of these coastal zones, despite being approximate in nature and ignoring longshore transport (Hancock et al., 2006; Wang et al., 2007). The recent studies to investigate the dynamics of hypersalinity in the GBR have significant limitations. For instance, the 1D cross-shelf exchange and diffusion models neglect longshore gradient of the salinity of the bays in the coastal zones of the GBR (Hancock et al., 2006; Wang et al., 2007). In addition, the 2D vertical integrated-models (Andutta et al., 2011) ignore baroclinic force driving thermohaline circulation (Nunes and Lennon, 1986) even though they have successfully duplicated the spatial distribution of hypersalinity along the GBR coast lines (Andutta et al., 2011); these 2D models required considerable manipulation of model diffusion coefficients to produce results comparable to field data. Ideally, a 3D model description of the circulation is required in order to simulate the baroclinic forcing, the general wind and tidal barotropic flow especially along the shelf. Furthermore, this 3D-model can also simulate the influence of the Coriolis Effect on the baroclinic flow, i.e. the sub-surface hypersaline flow might be deviated by Coriolis Effect in the near-bottom layer as reported observationally (Lavı́n et al., 1998). This deviation of the salt transport cannot be described by 1D-model and 2D-model due to the absence of longshore and vertical aspects, respectively. The objectives of this study are to (a) simulate 3D-features of the dynamic of hypersalinity in the coastal zones of GBR, (b) investigate the density-driven circulation in this area and (c) calculate the flushing time of the hypersaline water in selected bays in association with the prevailing transports in the coastal waters (i.e. longshore and the cross-shore transports). In order to keep the computational cost of the 3D model reasonable, and due to limitations on the availability of coastal salinity data to be used for model validation, a subsection of the GBR lagoon was chosen. The area of interest for this numerical study is the shallow-water environment situated from Bowling Green Bay to Halifax Bay (Fig. 2b) of the central GBR region (Fig.2a). This area is in the dry tropics region of the GBR and hypersaline conditions are a regular feature of the dry season (Walker, 1981b; Wolanski and Jones, 1981b). 2. Physical description of the study area During the dry season (April to November), the central GBR (Fig.2a) is dominated by the south-east trade wind (Wolanski, 1982) with average evaporation rate over the ocean of 5 mm/day (Da Silva et al.,
Fig. 2. (a) The central GBR, and (b) one of coastal zones of the central GBR including Halifax Bay, Cleveland Bay and Bowling Green Bay; (•) represents oceanographic stations from routine measurements using Seabird SBE 19 (i.e. 19–21 September 2009, 18 September 2010, 10–11 September 2011, 7–9 September 2012 and 5–7 October 2013); (×) describes oceanographic stations of Wolanski and Jones (1981b); transects A, B and C represent cross-shelf transects of eastern Cape Cleveland, Cleveland Bay and western Magnetic Island, respectively for which model results are calculated.
1994; Gibson et al., 1999; Kalnay et al., 1996) and negligible precipitation (Wang et al., 2007). Semi-diurnal tides prevail in these coastal zones with the surface elevation ranging from 0.5 m to 3.8 m (Hamon, 1984; Lou, 1995). Coastal zone oceanography is affected by oceanic inflows from the Coral Sea (Brinkman et al., 2002) and this inflow is significantly affected by the complex bathymetry of this region (Brinkman et al., 2002; Wolanski, 1994). The residual circulation is affected by the oceanic inflow and also by the SE trade winds which produce a generally longshore transport near the coasts (Wolanski, 1994). Cleveland Bay and Bowling Green Bay in the central GBR (Fig. 2b) have been the focus of numerous surveys of coastal hypersalinity (Walker, 1981b, 1982; Wang et al., 2007; Wolanski, 1994; Wolanski et al., 1981). Furthermore, the routine salinity measurements for this coastal hypersalinity have also, recently, been conducted inside the Cleveland Bay (Fig. 2b). Geomorphologically, Cleveland and Bowling Green Bays are 25 km and 80 km wide, respectively and, relatively 15 m deep at their seaward edge (Fig. 2b) (Lou, 1995; Wolanski and Jones, 1981a).
3. Material and methods 3.1. The MOHID model The 3D-model used in this study was the MOHID model (www. mohid.com). MOHID is the 3D water modelling system developed by the Marine and Environmental Technology Research Centre (MARETEC) (Mateus and Neves, 2013; Mateus et al., 2012). MOHID has been used not only in Portugal (Cancino and Neves, 1999; Martins et al., 2001; Vaz et al., 2007) but also in other regional areas including the Ria de Vigo, Spain (Taboada et al., 1998), Western Europe margin (Coelho
G.G. Salamena et al. / Marine Pollution Bulletin 105 (2016) 277–285
et al., 1999), Japan (Mateus and Neves, 2013), Brazil (Deus et al., 2013) and the Tyrrhenian Sea (Janeiro et al., 2014). The Hydrodynamics Module of MOHID is a 3D-baroclinic model solving the shallow waters equations using the hydrostatic assumption and the Boussinesq approximation (Martins et al., 2001): ∂ui ¼ 0; ∂xi Zη ∂ui ∂ ui u j 1 ∂patm ρðηÞ ∂η g ∂ρ0 ¼− −g − dx3 þ ρ0 ∂xi ρ0 ∂xi ρ0 ∂xi ∂t ∂x j x3 ! ∂ ∂u þ ν i −2εijk Ω j uk ; ∂x j ∂x j
ð1Þ
ð2Þ
where ui are the velocity vector components in the Cartesian xi directions while η, ν and patm are the free surface elevation, the turbulent viscosity, and the atmospheric pressure, respectively. ρ and ρ′ are the water density and its anomaly, respectively (ρ = ρ′ + ρ0) while the Coriolis force (i.e. the Earth rotation, Ωj and the Kroneker operator, εi,j,k) is represented by the last term in Eq. (2). The seawater density is computed by the UNESCO equation of state (UNESCO, 1981) using input from the salinity (S) and temperature (T) values computed by the model. S and T are transported using the same methods used for the momentum equations (Martins et al., 2001). The MOHID model handles heat and mass fluxes at the oceanic-atmospheric boundary using its Coupling Water–Atmosphere module; the theoretical frameworks of this module follow Ohlmann et al. (2000) and Ohlmann and Siegel (2000). Numerically, the transport equations are discretised using the finite volume method using an Arakawa-C staggered grid (Arakawa, 1966) and solved by a semi-implicit Alternating Direction Implicit (ADI) algorithm calculating the change of water elevation and velocity evolution based on Leendertse (1967) and Abbott et al. (1973) (see Martins et al., (1998)). The vertical coordinate in the MOHID model is of a generic type which enables combination of several coordinate types (i.e. sigma, Cartesian, Isopycnic or Lagrangian) (Martins et al., 1998). This is achieved directly due to the finite volume method instead of using the coordinate transformations (Neves et al., 2000). The horizontal and vertical transports of temperature and salinity are computed explicitly and implicitly, respectively (Mateus et al., 2012). MOHID enables users to utilize one of several approaches of horizontal turbulent diffusion i.e. constant diffusion coefficient, Smagorinsky method (Smagorinsky, 1963) or bi-harmonic approach (Delhez and Deleersnijder, 2007). Furthermore, the baroclinic force is computed using a z-level technique for any vertical coordinate (i.e. sigma and Cartesian) by which the horizontal density gradient is calculated (Mateus et al., 2012). In addition, the bottom stress is also calculated implicitly as a part of boundary condition associated with the vertical diffusion term (Mateus et al., 2012). The MOHID model uses the GOTM (General Ocean Turbulence Model) in its code in order to solve the vertical turbulent kinetic equations (Baumert, 2005; Burchard, 2002; Ruiz-Villarreal et al., 2005). In the coupling of the GOTM to the MOHID model, users can choose either k-l model (Mellor and Yamada, 1974, 1982) or k-ε model (Rodi, 1980, 1987) which is widely used for turbulence modelling in marine waters (Canuto et al., 2001; Galperin et al., 1988; Kantha and Clayson, 1994; Luyten et al., 1996). The MOHID model uses the Lagrangian approach embedded in its Lagrangian transport module for flushing time calculations (Braunschweig et al., 2003; Pando et al., 2013). Technically, the Lagrangian tracers are initially established inside polygons (here called boxes) covering particular marine areas and the evolution of the quantity of the virtual tracers is monitored (Braunschweig et al., 2003; Kämpf et al., 2010).
279
The flushing time is estimated based on the decay of the tracers' fraction (Braunschweig et al., 2003) as shown by mðt Þ ¼ mo exp −t τ ;
ð3Þ
where m(t), mo, and τ are quantity of tracers at particular time t, the initial value of tracers and the flushing time, respectively (Braunschweig et al., 2003; Tartinville et al., 1997). The flushing time can be obtained once the ratio between m(t) and mo is approximately t ≈t, Wang et al. (2007), Deleersnijder et al. (2006)). 0.37 (τ ¼ − ln ð0:37Þ Another beneficial aspect of using boxes for Lagrangian application in the MOHID is to investigate the transfer of water particles from one box to another (Braunschweig et al., 2003). This allows one box to monitor the presence of Lagrangian tracers from other boxes inside it (Braunschweig et al., 2003); mathematically, this monitoring role is described by. f m;n ðt Þ ¼
V m;n ðt Þ ; V m;m ð0Þ
ð4Þ
where fm,n is the fraction of the Lagrangian tracers from box n inside box m, while Vm,n is volume of the Lagrangian tracers from box n inside box m at particular time t. Furthermore, Vm,m (0) is the total volume of Lagrangian tracers in box m at the initial time (t = 0) (Braunschweig et al., 2003). In addition, the total fraction of Lagrangian tracers from box n inside box m during a particular period, T can be estimated using Eq. (5) (Braunschweig et al., 2003).
F m;n ðT Þ ¼
1 T
ZT f m;n ðt Þ dt:
ð5Þ
0
Finally, it is worthwhile commenting on the potential limitations of the modelling methodology, in particular, the hydrostatic assumption. The hydrostatic assumption neglects the non-hydrostatic terms (vertical acceleration and the vertical components of the viscosity tensors). This assumption does not introduce significant distortion in the results for this particular case since the rate of change of the vertical process are three order of magnitude smaller than the horizontal ones (Chen, 2005). This assumption only fails for the conditions when vertical accelerations are comparable to the gravity acceleration e.g. at steep slopes (Horn et al., 2002; Weilbeer and Jankowski, 2000). Steep slopes enable an intense reflective process driven by strong tidal currents to cause significant vertical accelerations (Slinn and Riley, 1996; Wright, 1995). The geomorphology of the coastal hypersaline system of GBR is suitable for the application of the hydrostatic approximation. For instance, the gradual slope from coastal waters to the reef matrix in GBR and a moderate tidal currents in the inshore waters (e.g. c.a. 50 cm/s in Cleveland Bay) are likely to provide a dissipative process overshadowing the non-hydrostatic signatures (e.g. internal waves) (Jing and Ridd, 1997; Lou, 1995; Wright, 1995). In contrast, the hydrostatic approximation may fail at the shelf edge of GBR where steeper sloping terrain and strong mesoscale eddies behind reefs exist (Wolanski et al., 1996). These offshore waters of the GBR are not our main interest in this work as the hypersaline water occupies the inshore waters. 3.2. Model setup Model configuration for the area of interest was 3D-baroclinic with 20 vertical sigma layers uniformly spaced. This study used the k-ε model for the turbulence closure technique based on Canuto et al. (2001). For simplicity, the horizontal turbulent diffusion coefficient (Kh) was assumed to be equal to horizontal turbulent viscosity with the constant value of 20 m2/s. The initial surface salinity was 35.4‰ following Andutta et al. (2011) while the initial temperature was 25 °C based on observation. The open boundary conditions used were
280
G.G. Salamena et al. / Marine Pollution Bulletin 105 (2016) 277–285
winds and tidal data from Cape Ferguson near Cape Cleveland while the constant offshore boundary of salinity is 35.4‰, based on observations. The time periods of interest were between August and October (Andutta et al., 2011) corresponding to the dryest period of the year, and between mid November and December (the early of summer just before the rainfall starts to exist). The latter simulation time was used to investigate the dynamics of hypersalinity due to the increase of temperature. The processes considered in the MOHID's Coupling Water–Atmosphere module in this work were wind stress, heat and mass fluxes. The source of salinity at the surface was generated indirectly by specifying the atmosphere variable (air temperature, relative humidity, and wind). The evaporation rate over the GBR was set up to be constant (5 mm/day) via computation of latent heat flux in the module. This evaporation value is chosen according to several studies (Da Silva et al., 1994; Gibson et al., 1999; Kalnay et al., 1996). The bathymetric baseline data in this current work was downloaded from a high-resolution depth model for the Great Barrier Reef and Coral Sea (www.deepreef.org/projects/48-depth-model-gbr-project.html) with the resolution of 100 m. This data was used to produce the 1.1 km depth grid which was used in this application of the MOHID. For physical forcing, tides and the prevailing SE trade winds were used from observation station at Cape Ferguson, near Cape Cleveland and the dataset are available online from the Bureau of Meteorology (BOM), Australia (http://www.bom.gov.au). The atmospheric data (e.g. air temperature, atmospheric pressure and relative humidity) is also from the BOM's observation station. In terms of data usage for model validation, due to the absence of in-situ oceanographic measurements offshore from Cleveland Bay, Halifax Bay and Bowling Green Bay, the monthly average (2000–2014) of the sea surface temperature (SST) from the satellite Terra MODIS (4 km) (http://oceancolor.gsfc.nasa.gov/ cms/) was used. The oceanic inflow from the Coral Sea was ignored in this study despite its role in affecting the flushing process of the coastal embayments in the GBR (Andutta et al., 2013). The significant SE trade winds during the dry season prevent the intrusion of the oceanic inflow to cross the offshore reef matrix and thus, drive it to be transported seaward instead of entering the coastal hypersaline waters (Andutta et al., 2013). The contiguous boxes over the domain (Fig. 3) were utilized to apply the Lagrangian tracer-filled boxes following Braunschweig et al. (2003). The virtual tracers were released in Halifax Bay (Box 1: 320,868 tracers), Cleveland Bay (Box 2: 590,325 tracers) and Bowling Green Bay (Box 3: 102,593 tracers) and then monitored inside these embayments in order to calculate the flushing time. The different particle numbers inside these boxes are related to their volume while the effective volume of particles is uniformly 70,000 m3. To track the presence of the water
Fig. 3. Configuration of boxes for flushing time calculation and monitoring Lagrangian tracers in the study area; due to the initialization of the hypersalinity on the onshore waters, the virtual tracers were released only in Boxes 1, 2 and 3.
fraction from the coastal hypersaline embayments (Boxes 1, 2 and 3) transported longshore and cross-shelf, the monitoring role of all boxes was enabled. 3.3. The salt-exchange model The flushing time of Cleveland Bay was also calculated using the simple exchange model presented by Wang et al. (2007). This considers the salt balance between the exchange of water across the inshore-offshore salinity gradient and the evaporation regardless of the details of the flushing processes (i.e. horizontal turbulent diffusion or thermohaline circulation). The mathematical formula of this method for the time scale of the flushing process is. τ ¼ hðSI −SO Þ=ðE SO Þ;
ð6Þ
where h, SI and SO are the average depth of the coastal embayment, depth-averaged salinity in the inshore and offshore, respectively while E is the evaporation rate. 4. Results 4.1. Model validation Comparison of observed and model results for the dry season shows good agreement for both salinity and the cross-shelf gradient of temperature (Figs. 4 and 5). For instance, the model replicates the salinity gradient of the cross-shelf section of the observations (Fig. 4a and b). However, the model overestimates the absolute values of the SST satellite data (RMSE = 2.10 °C) but agrees well with the cross-shelf temperature gradient (Fig. 5). The salinity and the temperature gradients are arguably more important than the absolute values of the parameters due to the role of the onshore-offshore density gradient driving the density-driven circulation (Linden and Simpson, 1986; Wang et al., 2007). Therefore, despite the model–observation differences in the absolute value of coastal water temperature, the similar onshore–offshore density gradient between the model and the observations indicate the reliability of the model to investigate the density-driven circulation.
Fig. 4. (a) The comparison of the predicted salinity (averaged September value) using the 3D model to the observed data on September 2009 from oceanographic stations inside the Cleveland Bay and (b) model validation on the transect for the series of measurements in Sep.–Oct. 1979 from Wolanski and Jones (1981b).
G.G. Salamena et al. / Marine Pollution Bulletin 105 (2016) 277–285
281
Fig. 5. The comparison of normalized SST (SSTnorm = SST − SSTmean) between the average SST from model and the monthly climatology data of SST (2000–2014) on August from Terra MODIS 4 km resolution (source: http://oceancolor.gsfc.nasa.gov/cgi/l3); the data processing of the satellite data used SeaDAS (http://seadas.gsfc.nasa.gov).
4.2. Dynamics of the GBR's coastal hypersaline system
Fig. 7. The distribution of the average surface-bottom predicted salinity difference (August to October) in practical salinity units. Positive values represent surface salinity that is higher than bottom salinity.
The spatial distribution of the mean surface salinity from model calculations shows the hypersaline waters occupying the near-shore waters with the maximum value of 36.06‰ inside the Cleveland bay (Fig. 6). This agrees with the observation of hypersalinity in this bay from previous studies (Walker, 1981b; Wang et al., 2007). The establishment of the higher salinity water in the surface layer during the dry season can be illustrated by the small positive values from the distribution of the surface–near bottom salinity difference (Fig. 7). In general, these small positive values are predominant inside Halifax, Cleveland and the Bowling Green Bays due to the evaporation over the shallow waters (Heggie and Skyring, 1999; Meshal et al., 1984; Wolanski, 1986). In contrast, large negative values are present offshore where the inshore water have moved underneath the lower salinity surface layer (Fig. 7). The deeper waters play an important role in enabling significant thermohaline circulation. For instance, the existence of negative values in the offshore waters (Fig. 7), with the average depth of 25 m, indicates seaward movements of saltier water near-bed. Furthermore, these negative values are more considerable in the deep inshore waters including the eastern side of Cape Cleveland (c.a. 15 m depth) and the western side the Magnetic Island (c.a. 24 m depth) (Fig. 7). Moreover, the movement of subsurface high salinity water is most evident from the residual currents in the near-bottom layer (Fig. 8b) with the thermohaline currents reaching around 0.04 m/s, 0.08 m/s and 0.06 m/s for the
Fig. 6. Mean sea surface salinity distribution of model prediction between September and October in practical salinity units.
Fig. 8. (a) The surface and (b) the near-bottom residual currents integrated over the dry season (1 August to 31 October 2010); (c) the residual currents at the near-bottom layer integrated over the early summer season (18 November to 31 December 2010).
282
G.G. Salamena et al. / Marine Pollution Bulletin 105 (2016) 277–285
offshore waters of both Cleveland Bay and Halifax Bay, close to Cape Cleveland, and the western side of the Magnetic Island, respectively (Fig. 8b). In contrast, the north-westerly longshore transport dominates the net transport of the surface layer (Fig. 7a). In addition, the presence of a bottom saline tongue produced by thermohaline currents in transect A, B, C (Fig. 2b) further supports the evidence that the deeper bathymetry in these areas is important to the thermohaline circulation (Fig. 9a, b and c). The relatively intense density-driven circulation in the deeper waters is likely to occur due to the increase of depth minimizing the influence of the wind-driven vertical turbulence which can destroy the stratification (Dufois et al., 2008; Nepf and Geyer, 1996). This leads to the establishment of thermohaline currents (Lavı́n et al., 1998; Lennon et al., 1987a; Nepf and Geyer, 1996; Nunes and Lennon, 1987). In spite of the importance of the density-driven circulation during the dry season, the simulation in the early of summer (late of November to December) indicates that rising temperatures in the coastal waters mitigates the effect of rising salinity on the water density. For instance, the vertical salinity profile of transect A (Fig. 10b) and the net transport in the near-bottom layer (Fig. 8c) for the early summer imply considerably reduced thermohaline circulation compared to those during the dry winter (Figs. 8b and 9a). The higher SST occupies the inshore waters which also has high salinity (Fig. 10a). 4.3. Flushing time calculation The timescale of flushing process of hypersaline waters inside Halifax, Cleveland and Bowling Green Bays during the dry season can be determined by calculating the time for Lagrangian tracers that were initially inside Boxes 1, 2 and 3 (Fig. 3) to reduce to 37% of the original number. The flushing time of these coastal hypersaline waters was found to be 17 to 20 days (Table 1).
Fig. 10. (a) The cross-shelf profile of mean sea surface temperature (°C) and of mean sea surface salinity between mid-November and 31 December 2010; (b) the cross-section profiles of average salinity (from mid-November to 31 December 2010) for transect A.
The flushing time scale was also calculated using the salt exchange model (Eq. (6)) by utilizing the salinity data (September 2009) in the offshore and the inshore waters of Cleveland Bay (E = 0.005 ± 0.00125 m/day based on Wang et al. (2007); h = 8 ± 0.5 m; SI = 36.0 ± 0.10‰ and SO = 35.65 ± 0.10 ‰). The flushing time of Cleveland Bay using the exchange model is 17 ± 3 days, which illustrates an agreement with the 3D model calculation despite the great simplicity of the exchange model. 4.4. Trajectory of the hypersalinity transport The drift of the coastal hypersaline waters along and across the shelf can be gauged by considering the movement of Lagrangian drifters originally located in Halifax Bay (Box 1), Cleveland Bay (Box 2) and Bowling Green Bay (Box 3), see Table 2. This table shows the 30-day average of the volume of water in a particular box M that was originally sourced from Boxes 1, 2 and 3. Table 2 shows that the longshore transport to the north-west is very important with, for example, about 41% of the 30 day-averaged water volume in Box 1 (Halifax Bay) being originally sourced from Box 2 (Cleveland Bay). Similarly, 8% of the 30 day averaged water volume in Box 1 comes from Bowling Green Bay (Box 3). Moreover, about 18% of the 30 day averaged water volume in Cleveland Bay (Box 2), was sourced from Bowling Green Bay.
Table 1 The flushing time of Halifax Bay (Box 1), Cleveland Bay (Box 2) and Bowling Green Bay (Box 3).
Fig. 9. The cross-section profiles of dry season average salinity at (a) Cape Cleveland (transect A), (b) at the near western side of the Magnetic Island (transect C) and (c) Cleveland Bay (transect B). The locations of these cross-sections are found in Fig. 2b.
Boxes
Flushing time (days)
Box 1 Box 2 Box 3
17 20 17
G.G. Salamena et al. / Marine Pollution Bulletin 105 (2016) 277–285 Table 2 The 30-day averaged percentage of the water in Box M that was originally sourced from Box N (see Eq. (5)). Source box N
Box 1 (Halifax Bay)
Box 2 (Cleveland Bay)
Box 3 (Bowling Green Bay)
45% 0.0% 0.0% 0.0% 0.0% 0.0% 0.1% 5.0% 0.6% 0.0% 0.0% 0.0% 0.0% 0.0% 13.4% 1.7% 0.4%
41% 53.5% 0.0% 0.0% 0.0% 0.0% 7.0% 5.4% 1.6% 1.1% 0.0% 0.0% 0.0% 0.0% 0.0% 0.5% 0.2%
8% 18% 50% 0.0% 0.0% 6.5% 6.7% 0.7% 0.5% 2.6% 0.3% 0.0% 0.8% 0.0% 0.0% 0.0% 0.0%
Receiving box M Box 1 Box 2 Box 3 Box 4 Box 5 Box 6 Box 7 Box 8 Box 9 Box 10 Box 11 Box 12 Box 13 Box 14 Box 15 Box 16 Box 17
The cross-shelf transport from the bays to the adjacent offshore boxes was more moderate. For example, 5% of the 30 day average water volume in Box 8 was sourced from Box 1 (Halifax Bay) which is immediately offshore of Box 8. Similar small percentages are also seen for the drift of water from Cleveland Bay (Box 2) to the adjacent Box 7, and from Bowling Green Bay to Box 6, i.e. 7.0% and 6.7%, respectively. 5. Discussion The cross-shelf exchange processes in the coastal system of the GBR in the dry season is probably dominated by horizontal turbulent diffusion and the density-driven circulation which are associated with the formation of vertical salinity (Nepf and Geyer, 1996; Nunes and Lennon, 1987; Odd and Rodger, 1985; Vaz et al., 1989). For instance, the horizontal turbulent diffusion is likely more important in the wellmixed state than when the density-driven circulation is well developed (Vaz et al., 1989). As a result, this horizontal turbulent diffusion may rule the exchange process inside the shallow coastal embayments (e.g. Cleveland Bay, Bowling Green Bay and Halifax Bay). In contrast, the thermohaline circulation is more likely to dominate the cross-shelf transport when the stratification exists (Nunes and Lennon, 1987; Odd and Rodger, 1985), particularly in regions where shallow water embayments are close to regions of deep bathymetry e.g. along the transect A and C (Fig. 9a and b) and in offshore waters (Figs. 8b, 9c). The role of deep inshore waters along with the wind-driven longshore currents are crucial to the formation of the thermohaline circulation. The hypersaline waters drift longshore in a north-westerly direction due to the south-easterly trade winds and may eventually pass capes or headlands such as Cape Cleveland, where the water is considerably deeper (ca. 15 m). Here, a pronounced thermohaline circulation can occur. The inshore–offshore temperature gradient is also likely to affect the strength of the thermohaline circulation. Walker (1982) argued that the warmer temperature in the inshore waters can reduce the density of the hypersaline water in the embayments decreasing the tendency of the saltier waters to sink and thus reduces the density-driven circulation, significantly. As a result, the density-driven circulation of the hypersaline waters in the GBR's coastal waters is likely to be more significant during the dry winter season due to the relatively cold water occupying the inshore waters (Fig. 5) rather than during the early summer when the inshore SST is more than the offshore SST (Walker, 1981a). The flushing times (~ 2–3 weeks) of the GBR's coastal hypersaline embayments calculated by both the Lagrangian tracers and the simple salt-exchange model agree with previous studies. For instance, the
283
estimated flushing time using virtual tracers from 2D-model during dry season (Andutta et al., 2013) is around 11 days for other coastal hypersaline embayments with the area roughly similar with the Cleveland Bay (flushing time of Box B3 and B4 in Andutta et al. (2013)). Furthermore, this water timescale of flushing process in this coastal region was reported to be less than a month using the drifter trajectory approach (Choukroun et al., 2010), and the SST as a proxy of cross-shelf turbulent diffusion (Mao and Ridd, 2015). The 3D-model enables the significant thermohaline circulation in the deeper inshore waters of the GBR to be reported for the first time by this study. The intense thermohaline circulation exists due to the presence of high salinity at the deeper inshore waters driven by the wind-driven longshore currents from the regions of high salinity (e.g. Bowling Green Bay and Cleveland Bay). The 2D models of Andutta et al. (2011) can model the longshore transports of hypersalinity, but are unable to capture the seaward movement of subsurface hypersaline plume at the deeper inshore water. Moreover, the 1D models of Wang et al. (2007) cannot capture this physics as it ignores the important role of the longshore transport of hypersalinity modelled in Andutta et al. (2011). It is also notable that the transects of Wolanski et al. (1981) were well chosen as they showed the considerable density stratification (and thus thermohaline circulation) at a location which the model predicts should be most intense.
6. Concluding remarks The hypersaline waters occupy the shallow coastal embayments due to the high evaporation rates and almost zero freshwater input with the increase of salinity approximately of 1 psu above the offshore waters. The longshore currents are an important aspect for the flushing processes of the hypersalinity seaward as these longshore currents transport the coastal hypersaline waters into deeper inshore areas such as the capes and headlands (e.g. Cape Cleveland). In these deeper waters, a pronounced thermohaline circulation is predicted to exist to flush the hypersaline water offshore at velocities of up to 0.08 m/s. The flushing time of the hypersaline waters is predicted in the order of 2–3 weeks. However, due to the warmer temperature in the coastal embayments during the dry early summer, the inshore-offshore density gradient is predicted to be less than in the dry winter. This decreased inshore–offshore density difference can lead to reduced thermohaline circulation and slightly longer flushing time of the coastal hypersaline embayments. Under these circumstances, horizontal turbulent diffusion caused by the mesoscale turbulence, especially from wakes around headlands is likely to be the primary process causing flushing of coastal waters. It should be pointed out that there is presently no current meter data to directly support the prediction of enhanced offshore flow near the seabed in deeper inshore regions. However, the fact that the model is able to predict both the hypersalinity and temperature gradients with high accuracy gives some confidence in its veracity. The conclusion that high thermohaline circulation exists where hypersaline waters from the embayments meet deeper waters to the northwest can be applied to other areas of the GBR's coastal zones. For example, deep inshore waters close to Cape Upstart which is northwest of Abbot Bay, a region of hypersalinity (Wang et al., 2007) should also support a strong hypersaline offshore circulation. A useful focus of future work would be to directly measure the thermohaline circulation to improve the understanding of the dynamics of hypersaline waters in these coastal waters.
Acknowledgements This study was conducted possible by the Australian Awards Scholarship for MSc degree to Gerry Giliant Salamena in James Cook University, Australia. The kind assistance of J. Schlaefer for the tutorial of the MOHID model is acknowledged.
284
G.G. Salamena et al. / Marine Pollution Bulletin 105 (2016) 277–285
References Abbott, M., Damsgaard, A., Rodenhuis, G., 1973. System 21, Jupiter, a design system for two-dimensional nearly-horizontal flows. J. Hydraul. Res. 11, 1–28. Andutta, F.P., Ridd, P.V., Wolanski, E., 2011. Dynamics of hypersaline coastal waters in the Great Barrier Reef. Estuar. Coast. Shelf Sci. 94, 299–305. Andutta, F.P., Ridd, P.V., Wolanski, E., 2013. The age and the flushing time of the Great Barrier Reef waters. Cont. Shelf Res. 53, 11–19. Arakawa, A., 1966. Computational design for long-term numerical integration of the equations of fluid motion: two-dimensional incompressible flow. Part I. J. Comput. Phys. 1, 119–143. Baumert, H.Z., 2005. Marine Turbulence: Theories, Observations, and Models. Cambridge University Press. Braunschweig, F., Martins, F., Chambel, P., Neves, R., 2003. A methodology to estimate renewal time scales in estuaries: the Tagus estuary case. Ocean Dyn. 53, 137–145. Brinkman, R., Wolanski, E., Deleersnijder, E., McAllister, F., Skirving, W., 2002. Oceanic inflow from the Coral Sea into the Great Barrier reef. Estuar. Coast. Shelf Sci. 54, 655–668. Burchard, H., 2002. Applied Turbulence Modelling in Marine Waters. Springer, New York. Cancino, L., Neves, R., 1999. Hydrodynamic and sediment suspension modelling in estuarine systems: part I: description of the numerical models. J. Mar. Syst. 22, 105–116. Canuto, V., Howard, A., Cheng, Y., Dubovikov, M., 2001. Ocean turbulence. Part I: onepoint closure model-momentum and heat vertical diffusivities. J. Phys. Oceanogr. 31, 1413–1426. Chen, X., 2005. A comparison of hydrostatic and nonhydrostatic pressure components in seiche oscillations. Math. Comput. Model. 41, 887–902. Choukroun, S., Ridd, P.V., Brinkman, R., McKinna, L.I., 2010. On the surface circulation in the western Coral Sea and residence times in the Great Barrier Reef. J. Geophys. Res. Oceans 1978–2012, 115. Coelho, H., Neves, R., Leitão, P., Martins, H., Santos, A., 1999. The slope current along the western European margin: a numerical investigation. Bol. Inst. Esp. Oceanogr. 15, 61–72. Da Silva, A., Young, C., Levitus, S., 1994. Atlas of surface marine data 1994. Algorithms and Procedures, NOAA Atlas NESDIS 6 vol. 1. US Department of Commerce, NOAA, NESDIS, USA. de Silva Samarasinghe, J., 1989. Transient salt-wedges in a tidal gulf: a criterion for their formation. Estuar. Coast. Shelf Sci. 28, 129–148. de Silva Samarasinghe, J., Lennon, G., 1987. Hypersalinity, flushing and transient saltwedges in a tidal gulf—an inverse estuary. Estuar. Coast. Shelf Sci. 24, 483–498. Deleersnijder, E., Beckers, J.-M., Delhez, E.J., 2006. On the behaviour of the residence time at the bottom of the mixed layer. Environ. Fluid Mech. 6, 541–547. Delhez, E.J., Deleersnijder, E., 2007. Overshootings and spurious oscillations caused by biharmonic mixing. Ocean Model. 17, 183–198. Deus, R., Brito, D., Kenov, I.A., Lima, M., Costa, V., Medeiros, A., Neves, R., Alves, C., 2013. Three-dimensional model for analysis of spatial and temporal patterns of phytoplankton in Tucuruí reservoir, Pará, Brazil. Ecol. Model. 253, 28–43. Dufois, F., Garreau, P., Le Hir, P., Forget, P., 2008. Wave-and current-induced bottom shear stress distribution in the Gulf of Lions. Cont. Shelf Res. 28, 1920–1934. Galperin, B., Kantha, L., Hassid, S., Rosati, A., 1988. A quasi-equilibrium turbulent energy model for geophysical flows. J. Atmos. Sci. 45, 55–62. Gibson, J., Kallberg, P., Uppala, S., Hernandez, A., Nomura, A., Serrano, E., 1999. ECMWF reanalysis project report series, 1. ERA-15 Description, Version 2, ECMWF. UK, Reading. Gräwe, U., Wolff, J.-O., Ribbe, J., 2009. Mixing, hypersalinity and gradients in Hervey Bay, Australia. Ocean Dyn. 59, 643–658. Hamon, B.V., 1984. Diurnal and Semi-Diurnal Tides on the Great Barrier Reef. CSIRO Marine Laboratories, Australia. Hancock, G.J., Webster, I., Stieglitz, T.C., 2006. Horizontal mixing of Great Barrier Reef waters: offshore diffusivity determined from radium isotope distribution. J. Geophys. Res. Oceans 1978–2012, 111. Heggie, D., Skyring, G., 1999. Flushing of Australian estuaries, coastal lakes and embayments: an overview with biogeochemical commentary. AGSO J. Aust. Geol. Geophys. 17, 211–226. Horn, D., Imberger, J., Ivey, G., Redekopp, L., 2002. A weakly nonlinear model of long internal waves in closed basins. J. Fluid Mech. 467, 269–287. Janeiro, J., Zacharioudaki, A., Sarhadi, E., Neves, A., Martins, F., 2014. Enhancing the management response to oil spills in the Tuscany archipelago through operational modelling. Mar. Pollut. Bull. 85, 574–589. Jing, L., Ridd, P., 1997. Modelling of suspended sediment transport in coastal areas under waves and currents. Estuar. Coast. Shelf Sci. 45, 1–16. Kalnay, E., Kanamitsu, M., Kistler, R., Collins, W., Deaven, D., Gandin, L., Iredell, M., Saha, S., White, G., Woollen, J., 1996. The NCEP/NCAR 40-year reanalysis project. Bull. Am. Meteorol. Soc. 77, 437–471. Kämpf, J., Payne, N., Malthouse, P., 2010. Marine connectivity in a large inverse estuary. J. Coast. Res. 26, 1047–1056. Kantha, L.H., Clayson, C.A., 1994. An improved mixed layer model for geophysical applications. J. Geophys. Res. Oceans 1978–2012 (99), 25235–25266. Largier, J., Hollibaugh, J.T., Smith, S., 1997. Seasonally hypersaline estuaries in Mediterranean-climate regions. Estuar. Coast. Shelf Sci. 45, 789–797. Lavı́n, M., Godı́, V., Alvarez, L., 1998. Inverse-estuarine features of the upper Gulf of California. Estuar. Coast. Shelf Sci. 47, 769–795. Leendertse, J.J., 1967. Aspects of a Computational Model for Long-Period Water-Wave Propagation. Rand Corporation Santa Monica, California. Lennon, G., Bowers, D., Nunes, R., Scott, B., Ali, M., Boyle, J., Wenju, C., Herzfeld, M., Johansson, G., Nield, S., 1987a. Gravity currents and the release of salt from an inverse estuary. Nature 327, 695–697. Lennon, G.W., Bowers, D.G., Nunes, R.A., Scott, B.D., Ali, M., 1987b. Gravity currents and the release of Salt from an inverse estuary. Nature 327, 695.
Linden, P., Simpson, J., 1986. Gravity-driven flows in a turbulent fluid. J. Fluid Mech. 172, 481–497. Lou, J., 1995. Modelling of Hydrodynamic and Suspended Sediment Transport in Coastal Areas. James Cook University Of North QueenslandDept. of Physics, Townsville, Qld. Luyten, P.J., Deleersnijder, E., Ozer, J., Ruddick, K.G., 1996. Presentation of a family of turbulence closure models for stratified shallow water flows and preliminary application to the Rhine outflow region. Cont. Shelf Res. 16, 101–130. Mao, Y., Ridd, P.V., 2015. Sea surface temperature as a tracer to estimate cross-shelf turbulent diffusivity and flushing time in the Great Barrier Reef lagoon. J. Geophys. Res. Oceans 120, 4487–4502. Martins, F., Leitao, P., Silva, A., Neves, R., 2001. 3D modelling in the Sado estuary using a new generic vertical discretization approach. Oceanol. Acta 24, 51–62. Martins, F., Neves, R., Leitão, P.C., 1998. A three-dimensional hydrodynamic model with generic vertical coordinate. Hydroinformatics 1403e1410. Mateus, M., Neves, R., 2013. Ocean Modelling for Coastal Management — Case Studies With MOHID. IST Press, Lisboa, Portugal. Mateus, M., Riflet, G., Chambel, P., Fernandes, L., Fernandes, R., Juliano, M., Campuzano, F., Pablo, H.d., Neves, R., 2012. An operational model for the West Iberian coast: products and services. Ocean Sci. 8, 713–732. Mellor, G.L., Yamada, T., 1974. A hierarchy of turbulence closure models for planetary boundary layers. J. Atmos. Sci. 31, 1791–1806. Mellor, G.L., Yamada, T., 1982. Development of a turbulence closure model for geophysical fluid problems. Rev. Geophys. 20, 851–875. Meshal, A., Behairy, A., Osman, M., 1984. Evaporation from coastal and open waters of the central zone of the Red Sea. Atmosphere–Ocean 22, 369–378. Nahas, E.L., Pattiaratchi, C.B., Ivey, G.N., 2005. Processes controlling the position of frontal systems in Shark Bay, Western Australia. Estuar. Coast. Shelf Sci. 65, 463–474. Nepf, H.M., Geyer, W.R., 1996. Intratidal variations in stratification and mixing in the Hudson estuary. J. Geophys. Res. C Oceans 101, 12079–12086. Neves, R., Leitao, P., Braunschweig, F., Martins, F., Coelho, H., Santos, A., Miranda, R., 2000. The advantage of a generic coordinate approach for ocean modelling. Water Stud. 25– 34. Nunes, R., Lennon, G., 1986. Physical property distributions and seasonal trends in Spencer Gulf, South Australia: an inverse estuary. Mar. Freshw. Res. 37, 39–53. Nunes, R.A., Lennon, G.W., 1987. Episodic stratification and gravity currents in a marine environment of modulated turbulence. J. Geophys. Res. Oceans 1978–2012 (92), 5465–5480. Odd, N.M., Rodger, J., 1985. Hydraulic aspects of modelling water quality in deep estuaries and enclosed bays. In: Wilson, J.G., Halcrow, W. (Eds.), Estuarine Management and Quality Assessment. Springer, US, pp. 61–68. Ohlmann, J.C., Siegel, D.A., 2000. Ocean radiant heating. Part II: parameterizing solar radiation transmission through the upper ocean. J. Phys. Oceanogr. 30, 1849–1865. Ohlmann, J.C., Siegel, D.A., Mobley, C.D., 2000. Ocean radiant heating. Part I: optical influences. J. Phys. Oceanogr. 30, 1833–1848. Pando, S., Juliano, M., Garcia, R., Mendes, J., Thomsen, L., 2013. Application of a lagrangian transport model to organo-mineral aggregates within the Nazaré canyon. Biogeosciences 10, 4103–4115. Rodi, W., 1980. Turbulence models and their application in hydraulics—a state of the art review. NASA STI/Recon Technical Report A 81, p. 21395. Rodi, W., 1987. Examples of calculation methods for flow and mixing in stratified fluids. J. Geophys. Res. Oceans 1978–2012 (92), 5305–5328. Ruiz-Villarreal, M., Bolding, K., Burchard, H., Demirov, E., 2005. Coupling of the GOTM turbulence module to some threedimensional ocean models. In: Baumert, H.Z., Simpson, J.H., Sundermann, J. (Eds.), Marine Turbulence: Theories, Observations, and Models. Results of the CARTUM Projec. Cambridge University Press, pp. 225–237. Slinn, D.N., Riley, J.J., 1996. Turbulent mixing in the oceanic boundary layer caused by internal wave reflection from sloping terrain. Dyn. Atmos. Oceans 24, 51–62. Smagorinsky, J., 1963. General circulation experiments with the primitive equations: I. The basic experiment*. Mon. Weather Rev. 91, 99–164. Taboada, J., Prego, R., Ruiz-Villarreal, M., Gómez-Gesteira, M., Montero, P., Santos, A., Pérez-Villar, V., 1998. Evaluation of the seasonal variations in the residual circulation in the Rıa of Vigo (NW Spain) by means of a 3D baroclinic model. Estuar. Coast. Shelf Sci. 47, 661–670. Tartinville, B., Deleersnijder, E., Rancher, J., 1997. The water residence time in the Mururoa atoll lagoon: sensitivity analysis of a three-dimensional model. Coral Reefs 16, 193–203. UNESCO, 1981. Tenth report of the joint panel on oceanographic tables and standards. Technical Papers in Marine Science, p. 24. Vaz, N., Dias, J.M., Leitão, P.C., Nolasco, R., 2007. Application of the Mohid-2D model to a mesotidal temperate coastal lagoon. Comput. Geosci. 33, 1204–1209. Vaz, R.A.N., Lennon, G.W., de Silva Samarasinghe, J.R., 1989. The negative role of turbulence in estuarine mass transport. Estuar. Coast. Shelf Sci. 28, 361–377. Walker, T., 1981a. Annual temperature cycle in Cleveland Bay, Great Barrier Reef Province. Mar. Freshw. Res. 32, 987–991. Walker, T., 1981b. Seasonal salinty variations in Cleveland Bay, northern Queensland. Mar. Freshw. Res. 32, 143–149. Walker, T.A., 1982. Lack of evidence for evaporation-driven circulation in the Great Barrier Reef lagoon. Mar. Freshw. Res. 33, 717–722. Wang, Y., Ridd, P.V., Heron, M.L., Stieglitz, T.C., Orpin, A.R., 2007. Flushing time of solutes and pollutants in the central Great Barrier Reef lagoon, Australia. Mar. Freshw. Res. 58, 778–791. Weilbeer, H., Jankowski, J., 2000. A three-dimensional non-hydrostatic model for free surface flows—development, verification and limitations. Proc. of 6th International Conference Estuarine and Coastal Modeling, pp. 162–177. Wolanski, E., 1981. Physical oceanography of the Great Barrier Reef lagoon. Conference on Environmental Engineering 1981. National Commitee on Environmental Engineering, Institute of Engineers, Australia, pp. 9–13.
G.G. Salamena et al. / Marine Pollution Bulletin 105 (2016) 277–285 Wolanski, E., 1982. Low-level trade winds over the western Coral Sea. J. Appl. Meteorol. 21, 881–882. Wolanski, E., 1986. An evaporation-driven salinity maximum zone in Australian tropical estuaries. Estuar. Coast. Shelf Sci. 22, 415–424. Wolanski, E., 1994. Physical Oceanographic Processes of the Great Barrier Reef. CRC Press, Boca Raton, Fla. Wolanski, E., Asaeda, T., Tanaka, A., Deleersnijder, E., 1996. Three-dimensional island wakes in the field, laboratory experiments and numerical models. Cont. Shelf Res. 16, 1437–1452.
285
Wolanski, E., Jones, M., 1981a. Physical properties of Great Barrier Reef lagoon waters near Townsville. I. Effects of Burdekin River floods. Mar. Freshw. Res. 32, 305–319. Wolanski, E., Jones, M., Williams, W., 1981. Physical properties of the Great Barrier Reef lagoon waters near Townsville. II. Seasonal variations. Mar. Freshw. Res. 32, 321–334. Wolanski, E.E., Jones, M.M., 1981b. Biological, Chemical and Physical Observations in Inshore Waters of the Great Barrier Reef, North Queensland, 1979. Wright, L.D., 1995. Morphodynamics of Inner Continental Shelves. CRC Press, Boca Raton.
Contents lists available at ScienceDirect
Marine Pollution Bulletin journal homepage: www.elsevier.com/locate/marpolbul
Research paper
The density-driven circulation of the coastal hypersaline system of the Great Barrier Reef, Australia Gerry G. Salamena a,⁎, Flávio Martins b, Peter V. Ridd c a b c
College of Marine and Environmental Science, James Cook University, Townsville, Queensland 4811, Australia CIMA-EST/UAlg., Campus da Penha, P8000-117, Faro, Portugal Marine Geophysics Laboratory, College of Science Technology and Engineering, James Cook University, Townsville, Queensland 4811, Australia
a r t i c l e
i n f o
Article history: Received 16 September 2015 Received in revised form 30 January 2016 Accepted 4 February 2016 Available online 12 February 2016 Keywords: Density-driven circulation Hypersaline waters Great Barrier Reef Flushing time
a b s t r a c t The coastal hypersaline system of the Great Barrier Reef (GBR) in the dry season, was investigated for the first time using a 3D baroclinic model. In the shallow coastal embayments, salinity increases to c.a. 1‰ above typical offshore salinity (~35.4‰). This salinity increase is due to high evaporation rates and negligible freshwater input. The hypersalinity drifts longshore north-westward due to south-easterly trade winds and may eventually pass capes or headlands, e.g. Cape Cleveland, where the water is considerably deeper (c.a. 15 m). Here, a pronounced thermohaline circulation is predicted to occur which flushes the hypersalinity offshore at velocities of up to 0.08 m/s. Flushing time of the coastal embayments is around 2–3 weeks. During the dry season early summer, the thermohaline circulation reduces and therefore, flushing times are predicted to be slight longer due to the reduced onshore-offshore density gradient compared to that in the dry season winter period. © 2016 Elsevier Ltd. All rights reserved.
1. Introduction The existence of hypersaline waters in coastal zones of continental shelves is caused by the excess of evaporation over precipitation and river runoff (de Silva Samarasinghe and Lennon, 1987; Gräwe et al., 2009; Heggie and Skyring, 1999; Lavı́n et al., 1998; Wolanski, 1986). Hypersaline systems in continental shelves have been studied in numerous locations particularly around the dry continent of Australia, for example, in northern Australia (Wolanski, 1986), gulfs in the Southern Australia (de Silva Samarasinghe, 1989; de Silva Samarasinghe and Lennon, 1987; Nunes and Lennon, 1986, 1987), Hervey Bay, Australia (Gräwe et al., 2009) and coastal zones of Great Barrier Reef (GBR), Australia (Andutta et al., 2011; Wang et al., 2007; Wolanski, 1981). The extent of the hypersalinity depends upon the freshwater balance of the region which is highly associated with seasonal characteristics e.g. during the dry season summer (e.g. in Gulf of California, Lavı́n et al. (1998)) and the dry winter in the GBR (Walker, 1981b). One important hydrodynamical effect of the hypersaline waters is to produce a thermohaline circulation by which the saltier water masses will sink and be flushed out seaward along the sea-bed (Fig. 1) (Gräwe et al., 2009; Heggie and Skyring, 1999; Lennon et al., 1987b; Wolanski, 1986). As a result, this thermohaline circulation is an important oceanographic aspect in the hypersaline coastal waters of the continental shelf. ⁎ Corresponding author at: LIPI's Centre for Deep Sea Research, Indonesian Institute of Sciences (LIPI), Jl. Y. Syaranamual, Guru-Guru, Poka, Kota Ambon, Provinsi Maluku 97233, Indonesia. E-mail address: [email protected] (G.G. Salamena).
http://dx.doi.org/10.1016/j.marpolbul.2016.02.015 0025-326X/© 2016 Elsevier Ltd. All rights reserved.
The degree of hypersalinity on the continental shelf is not only influenced by the freshwater balance, but is also affected by the exchange of the coastal hypersaline waters with oceanic salinity from offshore (Wang et al., 2007). This exchange transport process is likely affected by turbulent diffusion, thermohaline circulation and large scale advection (Wang et al., 2007). Due to this exchange transport role, some authors have conducted studies connecting this transport process with the flushing time of the coastal hypersaline waters (de Silva Samarasinghe and Lennon, 1987; Hancock et al., 2006; Heggie and Skyring, 1999; Largier et al., 1997; Wang et al., 2007). The hypersaline system of the GBR shelf is different to most other hypersaline environments due to its aspect ratio. It is a continental shelf system 2000 km in the long-shelf direction and between 50 (further North) and 100 (further South) km across shelf. This contrasts with most other reported hypersaline systems which are bays, gulfs, or inverse estuaries and are smaller in the long-shelf direction than across shelf direction. Due to this morphological difference, along shore currents become important for the GBR (Andutta et al., 2011). In contrast, for waters in narrow bays, it is the cross shelf tidal currents which predominate e.g. Gulf of St. Vincent (de Silva Samarasinghe and Lennon, 1987), Gulf of California (Lavı́n et al., 1998), Shark Bay (Nahas et al., 2005) and San Diego Bay (Largier et al., 1997). Thus, along-shore currents of the GBR enables the hypersaline waters in the GBR to be transported alongshore from one coastal embayment to another (Andutta et al., 2011). There has been considerable works on residence or flushing times of the GBR waters due to the potential influence of residence time on pollutant build-up in the GBR lagoon (Andutta et al., 2013; Choukroun
278
G.G. Salamena et al. / Marine Pollution Bulletin 105 (2016) 277–285
Fig. 1. Schematic diagram of evaporation-driven circulation. The light grey indicates plume of hypersaline water. This figure is drawn based on Fig. 2d of Wang et al. (2007).
et al., 2010; Hancock et al., 2006; Mao and Ridd, 2015; Wang et al., 2007). For the inshore hypersaline system in the GBR, it is probable that a combination of longshore currents and cross-shelf transport including via the thermohaline circulation can significantly affect the exchange process between inshore and offshore waters. This contrasts with the situation in narrow bays and estuaries where the effects of longshore currents are negligible (Largier et al., 1997) and hence, the narrow waters are likely to have longer residence time (Kämpf et al., 2010). Recent studies investigating cross-shelf transport in the GBR's coastal hypersaline system, which used 1D-models of the diffusion of salt, indicate short residence times (a few weeks) of these coastal zones, despite being approximate in nature and ignoring longshore transport (Hancock et al., 2006; Wang et al., 2007). The recent studies to investigate the dynamics of hypersalinity in the GBR have significant limitations. For instance, the 1D cross-shelf exchange and diffusion models neglect longshore gradient of the salinity of the bays in the coastal zones of the GBR (Hancock et al., 2006; Wang et al., 2007). In addition, the 2D vertical integrated-models (Andutta et al., 2011) ignore baroclinic force driving thermohaline circulation (Nunes and Lennon, 1986) even though they have successfully duplicated the spatial distribution of hypersalinity along the GBR coast lines (Andutta et al., 2011); these 2D models required considerable manipulation of model diffusion coefficients to produce results comparable to field data. Ideally, a 3D model description of the circulation is required in order to simulate the baroclinic forcing, the general wind and tidal barotropic flow especially along the shelf. Furthermore, this 3D-model can also simulate the influence of the Coriolis Effect on the baroclinic flow, i.e. the sub-surface hypersaline flow might be deviated by Coriolis Effect in the near-bottom layer as reported observationally (Lavı́n et al., 1998). This deviation of the salt transport cannot be described by 1D-model and 2D-model due to the absence of longshore and vertical aspects, respectively. The objectives of this study are to (a) simulate 3D-features of the dynamic of hypersalinity in the coastal zones of GBR, (b) investigate the density-driven circulation in this area and (c) calculate the flushing time of the hypersaline water in selected bays in association with the prevailing transports in the coastal waters (i.e. longshore and the cross-shore transports). In order to keep the computational cost of the 3D model reasonable, and due to limitations on the availability of coastal salinity data to be used for model validation, a subsection of the GBR lagoon was chosen. The area of interest for this numerical study is the shallow-water environment situated from Bowling Green Bay to Halifax Bay (Fig. 2b) of the central GBR region (Fig.2a). This area is in the dry tropics region of the GBR and hypersaline conditions are a regular feature of the dry season (Walker, 1981b; Wolanski and Jones, 1981b). 2. Physical description of the study area During the dry season (April to November), the central GBR (Fig.2a) is dominated by the south-east trade wind (Wolanski, 1982) with average evaporation rate over the ocean of 5 mm/day (Da Silva et al.,
Fig. 2. (a) The central GBR, and (b) one of coastal zones of the central GBR including Halifax Bay, Cleveland Bay and Bowling Green Bay; (•) represents oceanographic stations from routine measurements using Seabird SBE 19 (i.e. 19–21 September 2009, 18 September 2010, 10–11 September 2011, 7–9 September 2012 and 5–7 October 2013); (×) describes oceanographic stations of Wolanski and Jones (1981b); transects A, B and C represent cross-shelf transects of eastern Cape Cleveland, Cleveland Bay and western Magnetic Island, respectively for which model results are calculated.
1994; Gibson et al., 1999; Kalnay et al., 1996) and negligible precipitation (Wang et al., 2007). Semi-diurnal tides prevail in these coastal zones with the surface elevation ranging from 0.5 m to 3.8 m (Hamon, 1984; Lou, 1995). Coastal zone oceanography is affected by oceanic inflows from the Coral Sea (Brinkman et al., 2002) and this inflow is significantly affected by the complex bathymetry of this region (Brinkman et al., 2002; Wolanski, 1994). The residual circulation is affected by the oceanic inflow and also by the SE trade winds which produce a generally longshore transport near the coasts (Wolanski, 1994). Cleveland Bay and Bowling Green Bay in the central GBR (Fig. 2b) have been the focus of numerous surveys of coastal hypersalinity (Walker, 1981b, 1982; Wang et al., 2007; Wolanski, 1994; Wolanski et al., 1981). Furthermore, the routine salinity measurements for this coastal hypersalinity have also, recently, been conducted inside the Cleveland Bay (Fig. 2b). Geomorphologically, Cleveland and Bowling Green Bays are 25 km and 80 km wide, respectively and, relatively 15 m deep at their seaward edge (Fig. 2b) (Lou, 1995; Wolanski and Jones, 1981a).
3. Material and methods 3.1. The MOHID model The 3D-model used in this study was the MOHID model (www. mohid.com). MOHID is the 3D water modelling system developed by the Marine and Environmental Technology Research Centre (MARETEC) (Mateus and Neves, 2013; Mateus et al., 2012). MOHID has been used not only in Portugal (Cancino and Neves, 1999; Martins et al., 2001; Vaz et al., 2007) but also in other regional areas including the Ria de Vigo, Spain (Taboada et al., 1998), Western Europe margin (Coelho
G.G. Salamena et al. / Marine Pollution Bulletin 105 (2016) 277–285
et al., 1999), Japan (Mateus and Neves, 2013), Brazil (Deus et al., 2013) and the Tyrrhenian Sea (Janeiro et al., 2014). The Hydrodynamics Module of MOHID is a 3D-baroclinic model solving the shallow waters equations using the hydrostatic assumption and the Boussinesq approximation (Martins et al., 2001): ∂ui ¼ 0; ∂xi Zη ∂ui ∂ ui u j 1 ∂patm ρðηÞ ∂η g ∂ρ0 ¼− −g − dx3 þ ρ0 ∂xi ρ0 ∂xi ρ0 ∂xi ∂t ∂x j x3 ! ∂ ∂u þ ν i −2εijk Ω j uk ; ∂x j ∂x j
ð1Þ
ð2Þ
where ui are the velocity vector components in the Cartesian xi directions while η, ν and patm are the free surface elevation, the turbulent viscosity, and the atmospheric pressure, respectively. ρ and ρ′ are the water density and its anomaly, respectively (ρ = ρ′ + ρ0) while the Coriolis force (i.e. the Earth rotation, Ωj and the Kroneker operator, εi,j,k) is represented by the last term in Eq. (2). The seawater density is computed by the UNESCO equation of state (UNESCO, 1981) using input from the salinity (S) and temperature (T) values computed by the model. S and T are transported using the same methods used for the momentum equations (Martins et al., 2001). The MOHID model handles heat and mass fluxes at the oceanic-atmospheric boundary using its Coupling Water–Atmosphere module; the theoretical frameworks of this module follow Ohlmann et al. (2000) and Ohlmann and Siegel (2000). Numerically, the transport equations are discretised using the finite volume method using an Arakawa-C staggered grid (Arakawa, 1966) and solved by a semi-implicit Alternating Direction Implicit (ADI) algorithm calculating the change of water elevation and velocity evolution based on Leendertse (1967) and Abbott et al. (1973) (see Martins et al., (1998)). The vertical coordinate in the MOHID model is of a generic type which enables combination of several coordinate types (i.e. sigma, Cartesian, Isopycnic or Lagrangian) (Martins et al., 1998). This is achieved directly due to the finite volume method instead of using the coordinate transformations (Neves et al., 2000). The horizontal and vertical transports of temperature and salinity are computed explicitly and implicitly, respectively (Mateus et al., 2012). MOHID enables users to utilize one of several approaches of horizontal turbulent diffusion i.e. constant diffusion coefficient, Smagorinsky method (Smagorinsky, 1963) or bi-harmonic approach (Delhez and Deleersnijder, 2007). Furthermore, the baroclinic force is computed using a z-level technique for any vertical coordinate (i.e. sigma and Cartesian) by which the horizontal density gradient is calculated (Mateus et al., 2012). In addition, the bottom stress is also calculated implicitly as a part of boundary condition associated with the vertical diffusion term (Mateus et al., 2012). The MOHID model uses the GOTM (General Ocean Turbulence Model) in its code in order to solve the vertical turbulent kinetic equations (Baumert, 2005; Burchard, 2002; Ruiz-Villarreal et al., 2005). In the coupling of the GOTM to the MOHID model, users can choose either k-l model (Mellor and Yamada, 1974, 1982) or k-ε model (Rodi, 1980, 1987) which is widely used for turbulence modelling in marine waters (Canuto et al., 2001; Galperin et al., 1988; Kantha and Clayson, 1994; Luyten et al., 1996). The MOHID model uses the Lagrangian approach embedded in its Lagrangian transport module for flushing time calculations (Braunschweig et al., 2003; Pando et al., 2013). Technically, the Lagrangian tracers are initially established inside polygons (here called boxes) covering particular marine areas and the evolution of the quantity of the virtual tracers is monitored (Braunschweig et al., 2003; Kämpf et al., 2010).
279
The flushing time is estimated based on the decay of the tracers' fraction (Braunschweig et al., 2003) as shown by mðt Þ ¼ mo exp −t τ ;
ð3Þ
where m(t), mo, and τ are quantity of tracers at particular time t, the initial value of tracers and the flushing time, respectively (Braunschweig et al., 2003; Tartinville et al., 1997). The flushing time can be obtained once the ratio between m(t) and mo is approximately t ≈t, Wang et al. (2007), Deleersnijder et al. (2006)). 0.37 (τ ¼ − ln ð0:37Þ Another beneficial aspect of using boxes for Lagrangian application in the MOHID is to investigate the transfer of water particles from one box to another (Braunschweig et al., 2003). This allows one box to monitor the presence of Lagrangian tracers from other boxes inside it (Braunschweig et al., 2003); mathematically, this monitoring role is described by. f m;n ðt Þ ¼
V m;n ðt Þ ; V m;m ð0Þ
ð4Þ
where fm,n is the fraction of the Lagrangian tracers from box n inside box m, while Vm,n is volume of the Lagrangian tracers from box n inside box m at particular time t. Furthermore, Vm,m (0) is the total volume of Lagrangian tracers in box m at the initial time (t = 0) (Braunschweig et al., 2003). In addition, the total fraction of Lagrangian tracers from box n inside box m during a particular period, T can be estimated using Eq. (5) (Braunschweig et al., 2003).
F m;n ðT Þ ¼
1 T
ZT f m;n ðt Þ dt:
ð5Þ
0
Finally, it is worthwhile commenting on the potential limitations of the modelling methodology, in particular, the hydrostatic assumption. The hydrostatic assumption neglects the non-hydrostatic terms (vertical acceleration and the vertical components of the viscosity tensors). This assumption does not introduce significant distortion in the results for this particular case since the rate of change of the vertical process are three order of magnitude smaller than the horizontal ones (Chen, 2005). This assumption only fails for the conditions when vertical accelerations are comparable to the gravity acceleration e.g. at steep slopes (Horn et al., 2002; Weilbeer and Jankowski, 2000). Steep slopes enable an intense reflective process driven by strong tidal currents to cause significant vertical accelerations (Slinn and Riley, 1996; Wright, 1995). The geomorphology of the coastal hypersaline system of GBR is suitable for the application of the hydrostatic approximation. For instance, the gradual slope from coastal waters to the reef matrix in GBR and a moderate tidal currents in the inshore waters (e.g. c.a. 50 cm/s in Cleveland Bay) are likely to provide a dissipative process overshadowing the non-hydrostatic signatures (e.g. internal waves) (Jing and Ridd, 1997; Lou, 1995; Wright, 1995). In contrast, the hydrostatic approximation may fail at the shelf edge of GBR where steeper sloping terrain and strong mesoscale eddies behind reefs exist (Wolanski et al., 1996). These offshore waters of the GBR are not our main interest in this work as the hypersaline water occupies the inshore waters. 3.2. Model setup Model configuration for the area of interest was 3D-baroclinic with 20 vertical sigma layers uniformly spaced. This study used the k-ε model for the turbulence closure technique based on Canuto et al. (2001). For simplicity, the horizontal turbulent diffusion coefficient (Kh) was assumed to be equal to horizontal turbulent viscosity with the constant value of 20 m2/s. The initial surface salinity was 35.4‰ following Andutta et al. (2011) while the initial temperature was 25 °C based on observation. The open boundary conditions used were
280
G.G. Salamena et al. / Marine Pollution Bulletin 105 (2016) 277–285
winds and tidal data from Cape Ferguson near Cape Cleveland while the constant offshore boundary of salinity is 35.4‰, based on observations. The time periods of interest were between August and October (Andutta et al., 2011) corresponding to the dryest period of the year, and between mid November and December (the early of summer just before the rainfall starts to exist). The latter simulation time was used to investigate the dynamics of hypersalinity due to the increase of temperature. The processes considered in the MOHID's Coupling Water–Atmosphere module in this work were wind stress, heat and mass fluxes. The source of salinity at the surface was generated indirectly by specifying the atmosphere variable (air temperature, relative humidity, and wind). The evaporation rate over the GBR was set up to be constant (5 mm/day) via computation of latent heat flux in the module. This evaporation value is chosen according to several studies (Da Silva et al., 1994; Gibson et al., 1999; Kalnay et al., 1996). The bathymetric baseline data in this current work was downloaded from a high-resolution depth model for the Great Barrier Reef and Coral Sea (www.deepreef.org/projects/48-depth-model-gbr-project.html) with the resolution of 100 m. This data was used to produce the 1.1 km depth grid which was used in this application of the MOHID. For physical forcing, tides and the prevailing SE trade winds were used from observation station at Cape Ferguson, near Cape Cleveland and the dataset are available online from the Bureau of Meteorology (BOM), Australia (http://www.bom.gov.au). The atmospheric data (e.g. air temperature, atmospheric pressure and relative humidity) is also from the BOM's observation station. In terms of data usage for model validation, due to the absence of in-situ oceanographic measurements offshore from Cleveland Bay, Halifax Bay and Bowling Green Bay, the monthly average (2000–2014) of the sea surface temperature (SST) from the satellite Terra MODIS (4 km) (http://oceancolor.gsfc.nasa.gov/ cms/) was used. The oceanic inflow from the Coral Sea was ignored in this study despite its role in affecting the flushing process of the coastal embayments in the GBR (Andutta et al., 2013). The significant SE trade winds during the dry season prevent the intrusion of the oceanic inflow to cross the offshore reef matrix and thus, drive it to be transported seaward instead of entering the coastal hypersaline waters (Andutta et al., 2013). The contiguous boxes over the domain (Fig. 3) were utilized to apply the Lagrangian tracer-filled boxes following Braunschweig et al. (2003). The virtual tracers were released in Halifax Bay (Box 1: 320,868 tracers), Cleveland Bay (Box 2: 590,325 tracers) and Bowling Green Bay (Box 3: 102,593 tracers) and then monitored inside these embayments in order to calculate the flushing time. The different particle numbers inside these boxes are related to their volume while the effective volume of particles is uniformly 70,000 m3. To track the presence of the water
Fig. 3. Configuration of boxes for flushing time calculation and monitoring Lagrangian tracers in the study area; due to the initialization of the hypersalinity on the onshore waters, the virtual tracers were released only in Boxes 1, 2 and 3.
fraction from the coastal hypersaline embayments (Boxes 1, 2 and 3) transported longshore and cross-shelf, the monitoring role of all boxes was enabled. 3.3. The salt-exchange model The flushing time of Cleveland Bay was also calculated using the simple exchange model presented by Wang et al. (2007). This considers the salt balance between the exchange of water across the inshore-offshore salinity gradient and the evaporation regardless of the details of the flushing processes (i.e. horizontal turbulent diffusion or thermohaline circulation). The mathematical formula of this method for the time scale of the flushing process is. τ ¼ hðSI −SO Þ=ðE SO Þ;
ð6Þ
where h, SI and SO are the average depth of the coastal embayment, depth-averaged salinity in the inshore and offshore, respectively while E is the evaporation rate. 4. Results 4.1. Model validation Comparison of observed and model results for the dry season shows good agreement for both salinity and the cross-shelf gradient of temperature (Figs. 4 and 5). For instance, the model replicates the salinity gradient of the cross-shelf section of the observations (Fig. 4a and b). However, the model overestimates the absolute values of the SST satellite data (RMSE = 2.10 °C) but agrees well with the cross-shelf temperature gradient (Fig. 5). The salinity and the temperature gradients are arguably more important than the absolute values of the parameters due to the role of the onshore-offshore density gradient driving the density-driven circulation (Linden and Simpson, 1986; Wang et al., 2007). Therefore, despite the model–observation differences in the absolute value of coastal water temperature, the similar onshore–offshore density gradient between the model and the observations indicate the reliability of the model to investigate the density-driven circulation.
Fig. 4. (a) The comparison of the predicted salinity (averaged September value) using the 3D model to the observed data on September 2009 from oceanographic stations inside the Cleveland Bay and (b) model validation on the transect for the series of measurements in Sep.–Oct. 1979 from Wolanski and Jones (1981b).
G.G. Salamena et al. / Marine Pollution Bulletin 105 (2016) 277–285
281
Fig. 5. The comparison of normalized SST (SSTnorm = SST − SSTmean) between the average SST from model and the monthly climatology data of SST (2000–2014) on August from Terra MODIS 4 km resolution (source: http://oceancolor.gsfc.nasa.gov/cgi/l3); the data processing of the satellite data used SeaDAS (http://seadas.gsfc.nasa.gov).
4.2. Dynamics of the GBR's coastal hypersaline system
Fig. 7. The distribution of the average surface-bottom predicted salinity difference (August to October) in practical salinity units. Positive values represent surface salinity that is higher than bottom salinity.
The spatial distribution of the mean surface salinity from model calculations shows the hypersaline waters occupying the near-shore waters with the maximum value of 36.06‰ inside the Cleveland bay (Fig. 6). This agrees with the observation of hypersalinity in this bay from previous studies (Walker, 1981b; Wang et al., 2007). The establishment of the higher salinity water in the surface layer during the dry season can be illustrated by the small positive values from the distribution of the surface–near bottom salinity difference (Fig. 7). In general, these small positive values are predominant inside Halifax, Cleveland and the Bowling Green Bays due to the evaporation over the shallow waters (Heggie and Skyring, 1999; Meshal et al., 1984; Wolanski, 1986). In contrast, large negative values are present offshore where the inshore water have moved underneath the lower salinity surface layer (Fig. 7). The deeper waters play an important role in enabling significant thermohaline circulation. For instance, the existence of negative values in the offshore waters (Fig. 7), with the average depth of 25 m, indicates seaward movements of saltier water near-bed. Furthermore, these negative values are more considerable in the deep inshore waters including the eastern side of Cape Cleveland (c.a. 15 m depth) and the western side the Magnetic Island (c.a. 24 m depth) (Fig. 7). Moreover, the movement of subsurface high salinity water is most evident from the residual currents in the near-bottom layer (Fig. 8b) with the thermohaline currents reaching around 0.04 m/s, 0.08 m/s and 0.06 m/s for the
Fig. 6. Mean sea surface salinity distribution of model prediction between September and October in practical salinity units.
Fig. 8. (a) The surface and (b) the near-bottom residual currents integrated over the dry season (1 August to 31 October 2010); (c) the residual currents at the near-bottom layer integrated over the early summer season (18 November to 31 December 2010).
282
G.G. Salamena et al. / Marine Pollution Bulletin 105 (2016) 277–285
offshore waters of both Cleveland Bay and Halifax Bay, close to Cape Cleveland, and the western side of the Magnetic Island, respectively (Fig. 8b). In contrast, the north-westerly longshore transport dominates the net transport of the surface layer (Fig. 7a). In addition, the presence of a bottom saline tongue produced by thermohaline currents in transect A, B, C (Fig. 2b) further supports the evidence that the deeper bathymetry in these areas is important to the thermohaline circulation (Fig. 9a, b and c). The relatively intense density-driven circulation in the deeper waters is likely to occur due to the increase of depth minimizing the influence of the wind-driven vertical turbulence which can destroy the stratification (Dufois et al., 2008; Nepf and Geyer, 1996). This leads to the establishment of thermohaline currents (Lavı́n et al., 1998; Lennon et al., 1987a; Nepf and Geyer, 1996; Nunes and Lennon, 1987). In spite of the importance of the density-driven circulation during the dry season, the simulation in the early of summer (late of November to December) indicates that rising temperatures in the coastal waters mitigates the effect of rising salinity on the water density. For instance, the vertical salinity profile of transect A (Fig. 10b) and the net transport in the near-bottom layer (Fig. 8c) for the early summer imply considerably reduced thermohaline circulation compared to those during the dry winter (Figs. 8b and 9a). The higher SST occupies the inshore waters which also has high salinity (Fig. 10a). 4.3. Flushing time calculation The timescale of flushing process of hypersaline waters inside Halifax, Cleveland and Bowling Green Bays during the dry season can be determined by calculating the time for Lagrangian tracers that were initially inside Boxes 1, 2 and 3 (Fig. 3) to reduce to 37% of the original number. The flushing time of these coastal hypersaline waters was found to be 17 to 20 days (Table 1).
Fig. 10. (a) The cross-shelf profile of mean sea surface temperature (°C) and of mean sea surface salinity between mid-November and 31 December 2010; (b) the cross-section profiles of average salinity (from mid-November to 31 December 2010) for transect A.
The flushing time scale was also calculated using the salt exchange model (Eq. (6)) by utilizing the salinity data (September 2009) in the offshore and the inshore waters of Cleveland Bay (E = 0.005 ± 0.00125 m/day based on Wang et al. (2007); h = 8 ± 0.5 m; SI = 36.0 ± 0.10‰ and SO = 35.65 ± 0.10 ‰). The flushing time of Cleveland Bay using the exchange model is 17 ± 3 days, which illustrates an agreement with the 3D model calculation despite the great simplicity of the exchange model. 4.4. Trajectory of the hypersalinity transport The drift of the coastal hypersaline waters along and across the shelf can be gauged by considering the movement of Lagrangian drifters originally located in Halifax Bay (Box 1), Cleveland Bay (Box 2) and Bowling Green Bay (Box 3), see Table 2. This table shows the 30-day average of the volume of water in a particular box M that was originally sourced from Boxes 1, 2 and 3. Table 2 shows that the longshore transport to the north-west is very important with, for example, about 41% of the 30 day-averaged water volume in Box 1 (Halifax Bay) being originally sourced from Box 2 (Cleveland Bay). Similarly, 8% of the 30 day averaged water volume in Box 1 comes from Bowling Green Bay (Box 3). Moreover, about 18% of the 30 day averaged water volume in Cleveland Bay (Box 2), was sourced from Bowling Green Bay.
Table 1 The flushing time of Halifax Bay (Box 1), Cleveland Bay (Box 2) and Bowling Green Bay (Box 3).
Fig. 9. The cross-section profiles of dry season average salinity at (a) Cape Cleveland (transect A), (b) at the near western side of the Magnetic Island (transect C) and (c) Cleveland Bay (transect B). The locations of these cross-sections are found in Fig. 2b.
Boxes
Flushing time (days)
Box 1 Box 2 Box 3
17 20 17
G.G. Salamena et al. / Marine Pollution Bulletin 105 (2016) 277–285 Table 2 The 30-day averaged percentage of the water in Box M that was originally sourced from Box N (see Eq. (5)). Source box N
Box 1 (Halifax Bay)
Box 2 (Cleveland Bay)
Box 3 (Bowling Green Bay)
45% 0.0% 0.0% 0.0% 0.0% 0.0% 0.1% 5.0% 0.6% 0.0% 0.0% 0.0% 0.0% 0.0% 13.4% 1.7% 0.4%
41% 53.5% 0.0% 0.0% 0.0% 0.0% 7.0% 5.4% 1.6% 1.1% 0.0% 0.0% 0.0% 0.0% 0.0% 0.5% 0.2%
8% 18% 50% 0.0% 0.0% 6.5% 6.7% 0.7% 0.5% 2.6% 0.3% 0.0% 0.8% 0.0% 0.0% 0.0% 0.0%
Receiving box M Box 1 Box 2 Box 3 Box 4 Box 5 Box 6 Box 7 Box 8 Box 9 Box 10 Box 11 Box 12 Box 13 Box 14 Box 15 Box 16 Box 17
The cross-shelf transport from the bays to the adjacent offshore boxes was more moderate. For example, 5% of the 30 day average water volume in Box 8 was sourced from Box 1 (Halifax Bay) which is immediately offshore of Box 8. Similar small percentages are also seen for the drift of water from Cleveland Bay (Box 2) to the adjacent Box 7, and from Bowling Green Bay to Box 6, i.e. 7.0% and 6.7%, respectively. 5. Discussion The cross-shelf exchange processes in the coastal system of the GBR in the dry season is probably dominated by horizontal turbulent diffusion and the density-driven circulation which are associated with the formation of vertical salinity (Nepf and Geyer, 1996; Nunes and Lennon, 1987; Odd and Rodger, 1985; Vaz et al., 1989). For instance, the horizontal turbulent diffusion is likely more important in the wellmixed state than when the density-driven circulation is well developed (Vaz et al., 1989). As a result, this horizontal turbulent diffusion may rule the exchange process inside the shallow coastal embayments (e.g. Cleveland Bay, Bowling Green Bay and Halifax Bay). In contrast, the thermohaline circulation is more likely to dominate the cross-shelf transport when the stratification exists (Nunes and Lennon, 1987; Odd and Rodger, 1985), particularly in regions where shallow water embayments are close to regions of deep bathymetry e.g. along the transect A and C (Fig. 9a and b) and in offshore waters (Figs. 8b, 9c). The role of deep inshore waters along with the wind-driven longshore currents are crucial to the formation of the thermohaline circulation. The hypersaline waters drift longshore in a north-westerly direction due to the south-easterly trade winds and may eventually pass capes or headlands such as Cape Cleveland, where the water is considerably deeper (ca. 15 m). Here, a pronounced thermohaline circulation can occur. The inshore–offshore temperature gradient is also likely to affect the strength of the thermohaline circulation. Walker (1982) argued that the warmer temperature in the inshore waters can reduce the density of the hypersaline water in the embayments decreasing the tendency of the saltier waters to sink and thus reduces the density-driven circulation, significantly. As a result, the density-driven circulation of the hypersaline waters in the GBR's coastal waters is likely to be more significant during the dry winter season due to the relatively cold water occupying the inshore waters (Fig. 5) rather than during the early summer when the inshore SST is more than the offshore SST (Walker, 1981a). The flushing times (~ 2–3 weeks) of the GBR's coastal hypersaline embayments calculated by both the Lagrangian tracers and the simple salt-exchange model agree with previous studies. For instance, the
283
estimated flushing time using virtual tracers from 2D-model during dry season (Andutta et al., 2013) is around 11 days for other coastal hypersaline embayments with the area roughly similar with the Cleveland Bay (flushing time of Box B3 and B4 in Andutta et al. (2013)). Furthermore, this water timescale of flushing process in this coastal region was reported to be less than a month using the drifter trajectory approach (Choukroun et al., 2010), and the SST as a proxy of cross-shelf turbulent diffusion (Mao and Ridd, 2015). The 3D-model enables the significant thermohaline circulation in the deeper inshore waters of the GBR to be reported for the first time by this study. The intense thermohaline circulation exists due to the presence of high salinity at the deeper inshore waters driven by the wind-driven longshore currents from the regions of high salinity (e.g. Bowling Green Bay and Cleveland Bay). The 2D models of Andutta et al. (2011) can model the longshore transports of hypersalinity, but are unable to capture the seaward movement of subsurface hypersaline plume at the deeper inshore water. Moreover, the 1D models of Wang et al. (2007) cannot capture this physics as it ignores the important role of the longshore transport of hypersalinity modelled in Andutta et al. (2011). It is also notable that the transects of Wolanski et al. (1981) were well chosen as they showed the considerable density stratification (and thus thermohaline circulation) at a location which the model predicts should be most intense.
6. Concluding remarks The hypersaline waters occupy the shallow coastal embayments due to the high evaporation rates and almost zero freshwater input with the increase of salinity approximately of 1 psu above the offshore waters. The longshore currents are an important aspect for the flushing processes of the hypersalinity seaward as these longshore currents transport the coastal hypersaline waters into deeper inshore areas such as the capes and headlands (e.g. Cape Cleveland). In these deeper waters, a pronounced thermohaline circulation is predicted to exist to flush the hypersaline water offshore at velocities of up to 0.08 m/s. The flushing time of the hypersaline waters is predicted in the order of 2–3 weeks. However, due to the warmer temperature in the coastal embayments during the dry early summer, the inshore-offshore density gradient is predicted to be less than in the dry winter. This decreased inshore–offshore density difference can lead to reduced thermohaline circulation and slightly longer flushing time of the coastal hypersaline embayments. Under these circumstances, horizontal turbulent diffusion caused by the mesoscale turbulence, especially from wakes around headlands is likely to be the primary process causing flushing of coastal waters. It should be pointed out that there is presently no current meter data to directly support the prediction of enhanced offshore flow near the seabed in deeper inshore regions. However, the fact that the model is able to predict both the hypersalinity and temperature gradients with high accuracy gives some confidence in its veracity. The conclusion that high thermohaline circulation exists where hypersaline waters from the embayments meet deeper waters to the northwest can be applied to other areas of the GBR's coastal zones. For example, deep inshore waters close to Cape Upstart which is northwest of Abbot Bay, a region of hypersalinity (Wang et al., 2007) should also support a strong hypersaline offshore circulation. A useful focus of future work would be to directly measure the thermohaline circulation to improve the understanding of the dynamics of hypersaline waters in these coastal waters.
Acknowledgements This study was conducted possible by the Australian Awards Scholarship for MSc degree to Gerry Giliant Salamena in James Cook University, Australia. The kind assistance of J. Schlaefer for the tutorial of the MOHID model is acknowledged.
284
G.G. Salamena et al. / Marine Pollution Bulletin 105 (2016) 277–285
References Abbott, M., Damsgaard, A., Rodenhuis, G., 1973. System 21, Jupiter, a design system for two-dimensional nearly-horizontal flows. J. Hydraul. Res. 11, 1–28. Andutta, F.P., Ridd, P.V., Wolanski, E., 2011. Dynamics of hypersaline coastal waters in the Great Barrier Reef. Estuar. Coast. Shelf Sci. 94, 299–305. Andutta, F.P., Ridd, P.V., Wolanski, E., 2013. The age and the flushing time of the Great Barrier Reef waters. Cont. Shelf Res. 53, 11–19. Arakawa, A., 1966. Computational design for long-term numerical integration of the equations of fluid motion: two-dimensional incompressible flow. Part I. J. Comput. Phys. 1, 119–143. Baumert, H.Z., 2005. Marine Turbulence: Theories, Observations, and Models. Cambridge University Press. Braunschweig, F., Martins, F., Chambel, P., Neves, R., 2003. A methodology to estimate renewal time scales in estuaries: the Tagus estuary case. Ocean Dyn. 53, 137–145. Brinkman, R., Wolanski, E., Deleersnijder, E., McAllister, F., Skirving, W., 2002. Oceanic inflow from the Coral Sea into the Great Barrier reef. Estuar. Coast. Shelf Sci. 54, 655–668. Burchard, H., 2002. Applied Turbulence Modelling in Marine Waters. Springer, New York. Cancino, L., Neves, R., 1999. Hydrodynamic and sediment suspension modelling in estuarine systems: part I: description of the numerical models. J. Mar. Syst. 22, 105–116. Canuto, V., Howard, A., Cheng, Y., Dubovikov, M., 2001. Ocean turbulence. Part I: onepoint closure model-momentum and heat vertical diffusivities. J. Phys. Oceanogr. 31, 1413–1426. Chen, X., 2005. A comparison of hydrostatic and nonhydrostatic pressure components in seiche oscillations. Math. Comput. Model. 41, 887–902. Choukroun, S., Ridd, P.V., Brinkman, R., McKinna, L.I., 2010. On the surface circulation in the western Coral Sea and residence times in the Great Barrier Reef. J. Geophys. Res. Oceans 1978–2012, 115. Coelho, H., Neves, R., Leitão, P., Martins, H., Santos, A., 1999. The slope current along the western European margin: a numerical investigation. Bol. Inst. Esp. Oceanogr. 15, 61–72. Da Silva, A., Young, C., Levitus, S., 1994. Atlas of surface marine data 1994. Algorithms and Procedures, NOAA Atlas NESDIS 6 vol. 1. US Department of Commerce, NOAA, NESDIS, USA. de Silva Samarasinghe, J., 1989. Transient salt-wedges in a tidal gulf: a criterion for their formation. Estuar. Coast. Shelf Sci. 28, 129–148. de Silva Samarasinghe, J., Lennon, G., 1987. Hypersalinity, flushing and transient saltwedges in a tidal gulf—an inverse estuary. Estuar. Coast. Shelf Sci. 24, 483–498. Deleersnijder, E., Beckers, J.-M., Delhez, E.J., 2006. On the behaviour of the residence time at the bottom of the mixed layer. Environ. Fluid Mech. 6, 541–547. Delhez, E.J., Deleersnijder, E., 2007. Overshootings and spurious oscillations caused by biharmonic mixing. Ocean Model. 17, 183–198. Deus, R., Brito, D., Kenov, I.A., Lima, M., Costa, V., Medeiros, A., Neves, R., Alves, C., 2013. Three-dimensional model for analysis of spatial and temporal patterns of phytoplankton in Tucuruí reservoir, Pará, Brazil. Ecol. Model. 253, 28–43. Dufois, F., Garreau, P., Le Hir, P., Forget, P., 2008. Wave-and current-induced bottom shear stress distribution in the Gulf of Lions. Cont. Shelf Res. 28, 1920–1934. Galperin, B., Kantha, L., Hassid, S., Rosati, A., 1988. A quasi-equilibrium turbulent energy model for geophysical flows. J. Atmos. Sci. 45, 55–62. Gibson, J., Kallberg, P., Uppala, S., Hernandez, A., Nomura, A., Serrano, E., 1999. ECMWF reanalysis project report series, 1. ERA-15 Description, Version 2, ECMWF. UK, Reading. Gräwe, U., Wolff, J.-O., Ribbe, J., 2009. Mixing, hypersalinity and gradients in Hervey Bay, Australia. Ocean Dyn. 59, 643–658. Hamon, B.V., 1984. Diurnal and Semi-Diurnal Tides on the Great Barrier Reef. CSIRO Marine Laboratories, Australia. Hancock, G.J., Webster, I., Stieglitz, T.C., 2006. Horizontal mixing of Great Barrier Reef waters: offshore diffusivity determined from radium isotope distribution. J. Geophys. Res. Oceans 1978–2012, 111. Heggie, D., Skyring, G., 1999. Flushing of Australian estuaries, coastal lakes and embayments: an overview with biogeochemical commentary. AGSO J. Aust. Geol. Geophys. 17, 211–226. Horn, D., Imberger, J., Ivey, G., Redekopp, L., 2002. A weakly nonlinear model of long internal waves in closed basins. J. Fluid Mech. 467, 269–287. Janeiro, J., Zacharioudaki, A., Sarhadi, E., Neves, A., Martins, F., 2014. Enhancing the management response to oil spills in the Tuscany archipelago through operational modelling. Mar. Pollut. Bull. 85, 574–589. Jing, L., Ridd, P., 1997. Modelling of suspended sediment transport in coastal areas under waves and currents. Estuar. Coast. Shelf Sci. 45, 1–16. Kalnay, E., Kanamitsu, M., Kistler, R., Collins, W., Deaven, D., Gandin, L., Iredell, M., Saha, S., White, G., Woollen, J., 1996. The NCEP/NCAR 40-year reanalysis project. Bull. Am. Meteorol. Soc. 77, 437–471. Kämpf, J., Payne, N., Malthouse, P., 2010. Marine connectivity in a large inverse estuary. J. Coast. Res. 26, 1047–1056. Kantha, L.H., Clayson, C.A., 1994. An improved mixed layer model for geophysical applications. J. Geophys. Res. Oceans 1978–2012 (99), 25235–25266. Largier, J., Hollibaugh, J.T., Smith, S., 1997. Seasonally hypersaline estuaries in Mediterranean-climate regions. Estuar. Coast. Shelf Sci. 45, 789–797. Lavı́n, M., Godı́, V., Alvarez, L., 1998. Inverse-estuarine features of the upper Gulf of California. Estuar. Coast. Shelf Sci. 47, 769–795. Leendertse, J.J., 1967. Aspects of a Computational Model for Long-Period Water-Wave Propagation. Rand Corporation Santa Monica, California. Lennon, G., Bowers, D., Nunes, R., Scott, B., Ali, M., Boyle, J., Wenju, C., Herzfeld, M., Johansson, G., Nield, S., 1987a. Gravity currents and the release of salt from an inverse estuary. Nature 327, 695–697. Lennon, G.W., Bowers, D.G., Nunes, R.A., Scott, B.D., Ali, M., 1987b. Gravity currents and the release of Salt from an inverse estuary. Nature 327, 695.
Linden, P., Simpson, J., 1986. Gravity-driven flows in a turbulent fluid. J. Fluid Mech. 172, 481–497. Lou, J., 1995. Modelling of Hydrodynamic and Suspended Sediment Transport in Coastal Areas. James Cook University Of North QueenslandDept. of Physics, Townsville, Qld. Luyten, P.J., Deleersnijder, E., Ozer, J., Ruddick, K.G., 1996. Presentation of a family of turbulence closure models for stratified shallow water flows and preliminary application to the Rhine outflow region. Cont. Shelf Res. 16, 101–130. Mao, Y., Ridd, P.V., 2015. Sea surface temperature as a tracer to estimate cross-shelf turbulent diffusivity and flushing time in the Great Barrier Reef lagoon. J. Geophys. Res. Oceans 120, 4487–4502. Martins, F., Leitao, P., Silva, A., Neves, R., 2001. 3D modelling in the Sado estuary using a new generic vertical discretization approach. Oceanol. Acta 24, 51–62. Martins, F., Neves, R., Leitão, P.C., 1998. A three-dimensional hydrodynamic model with generic vertical coordinate. Hydroinformatics 1403e1410. Mateus, M., Neves, R., 2013. Ocean Modelling for Coastal Management — Case Studies With MOHID. IST Press, Lisboa, Portugal. Mateus, M., Riflet, G., Chambel, P., Fernandes, L., Fernandes, R., Juliano, M., Campuzano, F., Pablo, H.d., Neves, R., 2012. An operational model for the West Iberian coast: products and services. Ocean Sci. 8, 713–732. Mellor, G.L., Yamada, T., 1974. A hierarchy of turbulence closure models for planetary boundary layers. J. Atmos. Sci. 31, 1791–1806. Mellor, G.L., Yamada, T., 1982. Development of a turbulence closure model for geophysical fluid problems. Rev. Geophys. 20, 851–875. Meshal, A., Behairy, A., Osman, M., 1984. Evaporation from coastal and open waters of the central zone of the Red Sea. Atmosphere–Ocean 22, 369–378. Nahas, E.L., Pattiaratchi, C.B., Ivey, G.N., 2005. Processes controlling the position of frontal systems in Shark Bay, Western Australia. Estuar. Coast. Shelf Sci. 65, 463–474. Nepf, H.M., Geyer, W.R., 1996. Intratidal variations in stratification and mixing in the Hudson estuary. J. Geophys. Res. C Oceans 101, 12079–12086. Neves, R., Leitao, P., Braunschweig, F., Martins, F., Coelho, H., Santos, A., Miranda, R., 2000. The advantage of a generic coordinate approach for ocean modelling. Water Stud. 25– 34. Nunes, R., Lennon, G., 1986. Physical property distributions and seasonal trends in Spencer Gulf, South Australia: an inverse estuary. Mar. Freshw. Res. 37, 39–53. Nunes, R.A., Lennon, G.W., 1987. Episodic stratification and gravity currents in a marine environment of modulated turbulence. J. Geophys. Res. Oceans 1978–2012 (92), 5465–5480. Odd, N.M., Rodger, J., 1985. Hydraulic aspects of modelling water quality in deep estuaries and enclosed bays. In: Wilson, J.G., Halcrow, W. (Eds.), Estuarine Management and Quality Assessment. Springer, US, pp. 61–68. Ohlmann, J.C., Siegel, D.A., 2000. Ocean radiant heating. Part II: parameterizing solar radiation transmission through the upper ocean. J. Phys. Oceanogr. 30, 1849–1865. Ohlmann, J.C., Siegel, D.A., Mobley, C.D., 2000. Ocean radiant heating. Part I: optical influences. J. Phys. Oceanogr. 30, 1833–1848. Pando, S., Juliano, M., Garcia, R., Mendes, J., Thomsen, L., 2013. Application of a lagrangian transport model to organo-mineral aggregates within the Nazaré canyon. Biogeosciences 10, 4103–4115. Rodi, W., 1980. Turbulence models and their application in hydraulics—a state of the art review. NASA STI/Recon Technical Report A 81, p. 21395. Rodi, W., 1987. Examples of calculation methods for flow and mixing in stratified fluids. J. Geophys. Res. Oceans 1978–2012 (92), 5305–5328. Ruiz-Villarreal, M., Bolding, K., Burchard, H., Demirov, E., 2005. Coupling of the GOTM turbulence module to some threedimensional ocean models. In: Baumert, H.Z., Simpson, J.H., Sundermann, J. (Eds.), Marine Turbulence: Theories, Observations, and Models. Results of the CARTUM Projec. Cambridge University Press, pp. 225–237. Slinn, D.N., Riley, J.J., 1996. Turbulent mixing in the oceanic boundary layer caused by internal wave reflection from sloping terrain. Dyn. Atmos. Oceans 24, 51–62. Smagorinsky, J., 1963. General circulation experiments with the primitive equations: I. The basic experiment*. Mon. Weather Rev. 91, 99–164. Taboada, J., Prego, R., Ruiz-Villarreal, M., Gómez-Gesteira, M., Montero, P., Santos, A., Pérez-Villar, V., 1998. Evaluation of the seasonal variations in the residual circulation in the Rıa of Vigo (NW Spain) by means of a 3D baroclinic model. Estuar. Coast. Shelf Sci. 47, 661–670. Tartinville, B., Deleersnijder, E., Rancher, J., 1997. The water residence time in the Mururoa atoll lagoon: sensitivity analysis of a three-dimensional model. Coral Reefs 16, 193–203. UNESCO, 1981. Tenth report of the joint panel on oceanographic tables and standards. Technical Papers in Marine Science, p. 24. Vaz, N., Dias, J.M., Leitão, P.C., Nolasco, R., 2007. Application of the Mohid-2D model to a mesotidal temperate coastal lagoon. Comput. Geosci. 33, 1204–1209. Vaz, R.A.N., Lennon, G.W., de Silva Samarasinghe, J.R., 1989. The negative role of turbulence in estuarine mass transport. Estuar. Coast. Shelf Sci. 28, 361–377. Walker, T., 1981a. Annual temperature cycle in Cleveland Bay, Great Barrier Reef Province. Mar. Freshw. Res. 32, 987–991. Walker, T., 1981b. Seasonal salinty variations in Cleveland Bay, northern Queensland. Mar. Freshw. Res. 32, 143–149. Walker, T.A., 1982. Lack of evidence for evaporation-driven circulation in the Great Barrier Reef lagoon. Mar. Freshw. Res. 33, 717–722. Wang, Y., Ridd, P.V., Heron, M.L., Stieglitz, T.C., Orpin, A.R., 2007. Flushing time of solutes and pollutants in the central Great Barrier Reef lagoon, Australia. Mar. Freshw. Res. 58, 778–791. Weilbeer, H., Jankowski, J., 2000. A three-dimensional non-hydrostatic model for free surface flows—development, verification and limitations. Proc. of 6th International Conference Estuarine and Coastal Modeling, pp. 162–177. Wolanski, E., 1981. Physical oceanography of the Great Barrier Reef lagoon. Conference on Environmental Engineering 1981. National Commitee on Environmental Engineering, Institute of Engineers, Australia, pp. 9–13.
G.G. Salamena et al. / Marine Pollution Bulletin 105 (2016) 277–285 Wolanski, E., 1982. Low-level trade winds over the western Coral Sea. J. Appl. Meteorol. 21, 881–882. Wolanski, E., 1986. An evaporation-driven salinity maximum zone in Australian tropical estuaries. Estuar. Coast. Shelf Sci. 22, 415–424. Wolanski, E., 1994. Physical Oceanographic Processes of the Great Barrier Reef. CRC Press, Boca Raton, Fla. Wolanski, E., Asaeda, T., Tanaka, A., Deleersnijder, E., 1996. Three-dimensional island wakes in the field, laboratory experiments and numerical models. Cont. Shelf Res. 16, 1437–1452.
285
Wolanski, E., Jones, M., 1981a. Physical properties of Great Barrier Reef lagoon waters near Townsville. I. Effects of Burdekin River floods. Mar. Freshw. Res. 32, 305–319. Wolanski, E., Jones, M., Williams, W., 1981. Physical properties of the Great Barrier Reef lagoon waters near Townsville. II. Seasonal variations. Mar. Freshw. Res. 32, 321–334. Wolanski, E.E., Jones, M.M., 1981b. Biological, Chemical and Physical Observations in Inshore Waters of the Great Barrier Reef, North Queensland, 1979. Wright, L.D., 1995. Morphodynamics of Inner Continental Shelves. CRC Press, Boca Raton.
