Numerical modelling of hydraulic control, solitary waves and primary instabilities in the Strait of Gibraltar
Introduction
The Strait of Gibraltar connects two major basins: the Northern Atlantic and the Mediterranean Sea, over which evaporation exceeds precipitation and river run-off. To compensate the resulting loss, exchanges of mass and salt are required through the strait. Fig. 1 illustrates the rather complex exchanges occurring there. Inflowing Atlantic water is less salty (salinity ) than the outflowing Mediterranean water (), and spreads as a surface layer in the Alboran Sea. The interface between the two water masses is distorted by undulations that are not precisely periodic with regard to the tidal cycle but exhibit regularity in some areas. One of the paper objectives is to better understand the small-scale processes that lead to the Atlantic and Mediterranean water masses transformation in the vicinity of the Strait of Gibraltar.
To further illustrate the exchange between the Northern Atlantic and the Mediterranean, a very simple steady-state model can be expressed as a system of two basic conservation equations.
Volume conservation is expressed as : while the conservation of salt requires: where is the Atlantic water volume flux (positive), is the Mediterranean water volume flux (negative), both localized in the Strait of Gibraltar, and is the space-averaged Evaporation minus Precipitation (and river runoff) water budget integrated over the whole Mediterranean Sea. is positive. () stands for Atlantic (Mediterranean) water mean salinity and (Bethoux, 1979). The water budget is positive in the Mediterranean due to excess evaporation that correspond to a yearly averaged loss of water of about 1 metre over the whole basin (Garrett et al., 1990).
A major dynamical feature in the Strait of Gibraltar is the so-called “flow criticality” usually characterized by the Froude number (): it compares the internal wave phase speed with a flow characteristic velocity. Several definitions of the non-dimensional Froude number can be found in the literature: it can notably be defined for each layer, resulting in a composite number for the whole water column, as in Farmer and Armi (1988) or in Sannino et al. (2009b).
A “subcritical” (respectively “supercritical”) regime lies in the range of small (respectively large) values of the Froude number (respectively ), with an intermediate “critical” regime for . The upstream propagation of internal waves is inhibited for supercritical flow so that a hydraulic control occurs at the transition from subcritical to supercritical flow; it persists during periods and within regions of large Froude numbers. As such, the hydraulic regime at a given point will vary in time according to substantial currents variations occurring along the tidal cycle. It is, for example, well established that large amplitude solitary waves in the Strait of Gibraltar can develop due to the hydraulic control at Camarinal Sill (Farmer and Armi, 1988), making it a crucial process to represent.
Several analytical models have been proposed to investigate the hydraulic control in the Gibraltar region (Bryden and Stommel, 1984, Farmer and Armi, 1986, Garrett et al., 1990). The hydraulic control usually occurs in these models at Camarinal Sill (CS), Espartel Sill (ES), and Tarifa Narrows (TN), although the modelled hydraulic control location and frequency vary according to the model refinement :
- 1.
Farmer and Armi (1986)’s two-layer model accounts for the strait geometry (depth and width), the exchanged volumes ( and ) and the salinity contrast (). This simple model is able to simulate two hydraulic controls: the first one located by the sill, the other in the TN contraction, defining “maximal exchange regime” (further details are given below).
- 2.
In a slightly more elaborated model, the inclusion of entrainment between the two layers and the subsequent interfacial layer introduction modify the left-hand terms of Eqs. (1), (2) with the introduction of horizontal and vertical transports in the interfacial layer (Bray et al., 1995). Critical conditions are changed within such two interfaces model which may support two baroclinic modes and new hydraulic controls (Sannino et al., 2009b).
- 3.
Considering a three-dimensional flow, the definition of the control needs to account for cross-strait variations such as the tilt of the density interface in the latitudinal direction. In the maximal exchange solution, control in TN may induce the detachment of the surface layer from the northern coast (Sannino et al., 2009b).
The hydraulic control effect within the strait is illustrated in Fig. 1. The flow is initially subcritical in the Strait; the propagation of internal waves is not hindered at the interface between Atlantic and Mediterranean waters (denoted “a” in Fig. 1); then the tidal flood in the vicinity of the Camarinal Sill becomes supercritical. In the supercritical to subcritical transition, downstream of the sill, a “hydraulic jump” (“b” in Fig. 1) may occur.
Hydraulic jumps are large-amplitude depressions in the regions where hydraulic controls occur. There, intense mixing between the Atlantic and Mediterranean waters takes place as observed by Wesson and Gregg (1994). Shear flow instabilities can develop in the hydraulic jump of the Camarinal Sill (denoted “c” in Fig. 1).
The release of hydraulic jumps generates large-amplitude, non-linear, nonhydrostatic Internal Solitary Waves (ISW) trains (denoted “d” in Fig. 1) (Farmer and Armi, 1988). As the barotropic tide is constrained by the bathymetry, large vertical velocities appear and induce energy transfer to several normal modes of internal waves. Some observations in the Strait of Gibraltar identify the largest ISW amplitude to the first baroclinic mode; for which vertical velocities have the same direction throughout the water column and all isopycnal surface displacements are in phase. The signature of Mode 2 waves (the vertical velocity profile exhibits one node) has also been observed in the region of Gibraltar strait (Farmer and Armi, 1988, Vázquez et al., 2006). The internal waves propagate at the interface of Mediterranean and Atlantic waters.
As the strait flow varies at various timescales during the year, some deviation is expected in the occurrence of the hydraulic control in the strait. This may have a wide impact since local flow conditions combined with the above two conservation equations (1), (2) determine the relation between the volume fluxes, the evaporation minus precipitation budget () and the salinity difference () (Bryden and Kinder, 1991). Practically, an “overmixed” solution corresponds to a minimal salinity difference and a maximal exchange of water mass in the strait: it would thus constrict the formation of Mediterranean waters and diapycnal mixing over the Mediterranean basin (Bryden and Stommel, 1984, Garrett et al., 1990). Moreover, the small-scale processes occurring in the strait itself can directly modify the local characteristics of Mediterranean waters (García-Lafuente et al., 2011, Naranjo et al., 2015) and Atlantic waters (Millot, 2014). This can affect their characteristics as they enter respectively in the North Atlantic sub-basin and in the Mediterranean Sea.
To study the flow dynamics in the strait in further details, more realistic numerical modelling is of great help. Early attempts used two-layer models (Brandt et al., 1996, Izquierdo et al., 2001). The increase of computational power led to 3D modelling (Sannino et al., 2004) with increasing vertical and horizontal resolution, explicitly addressing the tidal cycle and flow characteristics. More recently, even nonhydrostatic models have been used (Sánchez Garrido et al., 2011, Sannino et al., 2014) to explicitly represent the ISW. Other configurations include the Strait of Gibraltar into a Mediterranean circulation model (Soto-Navarro et al., 2015). In this case, the increased resolution locally in the strait (Naranjo et al., 2014) – or the nesting of high-resolution grids within a coarse resolved regional model (Sannino et al., 2009a) – shows a clear impact on Mediterranean stratification and improves the representation of convective events in the northwestern Mediterranean basin.
The coastal and regional ocean modelling community model (CROCO1 ) is based on a new nonhydrostatic and non-Boussinesq solver (Auclair et al., 2018) developed within the former ROMS kernel (Shchepetkin and McWilliams, 2005), for an optimal accuracy and cost efficiency. CROCO opens up new perspectives in terms of modelling of small-scale processes (Fox-Kemper et al., 2019, Lemarié et al., 2019). In this sense, the present study objectives are also numerical: we show that a new generation of nonhydrostatic ocean models can be used efficiently to simulate complex nonlinear, fine scale physics in a realistic but computationally-affordable configuration. The complete solution of Navier–Stokes equations is thus solved numerically for the very first time in a complex realistic regional configuration.
The present configuration of the Strait of Gibraltar is based on a classical lock-exchange initialization (Sannino et al., 2002). A 2D vertical section of the strait is adopted in order to reduce the number of parameters impacting the studied dynamics. This rather simple configuration is thus of weak computational cost and reduces the implementation burden; it allows to reach the horizontal and vertical scales of the largest turbulent structures observed in this area. In the strait, where most transverse dynamical feature are an order of magnitude weaker, our numerical approach is some kind of ersatz of a large-eddy simulation (LES2 ), for which at least the generation process of primary instabilities is correctly represented. However, LES is a 3D concept as the route to molecular dissipation differs in 2D and 3D turbulence. The present study is focused on the description of the largest primary instabilities in the Strait of Gibraltar; as well as providing order of magnitudes for explicit simulations of these dynamics. Along with these physical aims, the relevance of the chosen numerical methods is a major concern. A quantified impact of the largest turbulent structures on the water masses is out of the scope of what is presented hereafter: it would require a fully three dimensional LES (also achievable with the CROCO model), in complement with dedicated relevant experimental measurements.
In Section 2, we present an overview of CROCO equations and the implementation for the 2D lock-exchange experiment. We describe the implementation of the bathymetry profile, water masses, and the exchange and tidal flows. In Section 3, we analyse the physics of the 2D configuration, comparing the model solution to already published data (e.g., in Farmer and Armi (1988)). Emphasis is then made on the hydraulic control (Section 3.2), the hydraulic jump (Section 3.3) and the mode-1 and mode-2 ISW (non-linear internal trains of solitary waves) propagation (Section 3.4). Last, the sensitivity to the tidal forcing amplitude and to the numerical choices are analysed respectively in Sections 4.1 Tidal regime, 4.2 Nonhydrostatic balance and numerical factors, with a focus on the fine-scales dynamics listed in Fig. 1.
Section snippets
The numerical modelling system
The proposed numerical model of the Strait of Gibraltar simulates explicitly the fine-scale processes (from tens to hundreds of metres) discussed previously. This assumes that (i) a sufficient grid resolution is provided in the strait and (ii) a well-suited numerical kernel is used.
The nonhydrostatic (non-Boussinesq) CROCO version is chosen for its ability to allow the explicit representation of primary instabilities that cascade the kinetic energy injected at large scale down to the smaller
The reference simulation
The reference simulation presented previously is now evaluated thanks to the observational data from the Gibraltar Experiment (Farmer and Armi, 1988). We describe the hydraulic controls, the primary instabilities and the dynamics of the ISW in this reference simulation.
Sensitivity testing
The reference configuration presented in the previous section is based on several physical and numerical choices which are now investigated; mostly the impact of the forcing amplitude, momentum balance (hydrostatic approximation) and numerical parameters (spatial resolution, advection schemes).
Discussion and conclusion
The present study focuses on small-scale dynamics in the Strait of Gibraltar and on the capacity of a new split-explicit, free-surface, nonhydrostatic regional oceanic model (CROCO) to represent such dynamics. Both objectives were pursued in parallel and several seminal results are obtained.
The study confirms that the generation of large-amplitude mode-1 and mode-2 internal waves in the Strait of Gibraltar as well as the onset of stratified turbulence and its energy cascade can be simulated
CRediT authorship contribution statement
Margaux Hilt: Conceptualization, Software, Writing - original draft, Writing - review & editing, Visualization, Methodology, Investigation. Francis Auclair: Conceptualization, Software, Writing - review & editing, Funding acquisition, Supervision. Rachid Benshila: Software, Writing - review & editing. Lucie Bordois: Software, Writing - review & editing, Methodology. Xavier Capet: Software, Writing - review & editing. Laurent Debreu: Software, Writing - review & editing. Franck Dumas: Software,
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgements
This work was partly funded by the DGA “Etude Amont” Protevs driven by the Shom. It was granted access to the HPC ressources of CALMIP supercomputing center under the allocation P18017. We also gratefully thank the computer team of the Laboratoire d’Aérologie for its support. Margaux Hilt’s Ph.D. thesis was funded by a MESRI scholarship .
References (48)
- et al.
A non-hydrostatic non-Boussinesq algorithm for free-surface ocean modelling
Ocean Model.
(2018) - et al.
An improved weighted essentially non-oscillatory scheme for hyperbolic conservation laws
J. Comput. Phys.
(2008) - et al.
The boiling-water phenomena at Camarinal Sill, the Strait of Gibraltar
Deep-Sea Res. II
(2002) - et al.
Exchange through the Strait of Gibraltar
Prog. Oceanogr.
(1994) - et al.
Steady two-layer exchange through the Strait of Gibraltar
Deep Sea Res. A
(1991) - et al.
Two-way nesting in split-explicit ocean models: algorithms, implementation and validation
Ocean Model.
(2012) - et al.
The flow of Atlantic water through the Strait of Gibraltar
Prog. Oceanogr.
(1988) - et al.
Submesoscale dynamics in tropical instability waves
Ocean Model.
(2011) - et al.
Open boundary conditions for long-term integration of regional oceanic models
Ocean Model.
(2001) Heterogeneities of in- and out-flows in the Mediterranean Sea
Prog. Oceanogr.
(2014)
How much do tides affect the circulation of the Mediterranean Sea? From local processes in the Strait of Gibraltar to basin-scale effects
Prog. Oceanogr.
Mediterranean waters along and across the strait of gibraltar, characterization and zonal modification
Deep Sea Res. I
Combined wavelet and principal component analysis (WEof) of a scale-oriented model of coastal ocean gravity waves
Dyn. Atmos. Oceans
An eddy-permitting model of the Mediterranean Sea with two-way grid refinement at the Strait of Gibraltar
Ocean Model.
On effective resolution in ocean models
Ocean Model.
Budgets of the Mediterranean Sea. Their dependance on the local climate and on the characteristics of the Atlantic waters
Oceanol. Acta
Generation of bathymetric digital elevation models along French coasts: Coastal risk assessment
Hydro Int.
The effect of rotation on the surface inlow through the Strait of Gibraltar
J. Phys. Oceanogr.
Study of the generation and propagation of internal waves in the Strait of Gibraltar using a numerical model and synthetic aperture radar images of the European ERS 1 satellite
J. Geophys. Res.
The role of the interface in exchange through the Strait of Gibraltar
J. Geophys. Res.
Limiting processes that determine basic features of the circulation in the Mediterranean Sea
Oceanol. Acta
Tides in the Strait of Gibraltar
J. Geophys. Res.: Oceans
Global observations of large oceanic eddies
Geophys. Res. Lett.
Cited by (24)
Understanding tidal mixing at the Strait of Gibraltar: A high-resolution model approach
2023, Progress in OceanographyCirculation and stratification drivers during the summer season in the upwelling bay of Paracas (Peru): A modelling study
2023, Continental Shelf ResearchCitation Excerpt :A linear bottom friction (with a drag coefficient of 3.10−4 s−1) is used. Bulk formulae from Fairall et al. (2003) were used to compute atmospheric heat and momentum fluxes. A one-way offline nesting strategy based on the “ROMS2ROMS” downscaling (Mason et al., 2010) was set up to study the circulation in the bay.
Improved Gulf Stream separation through Brinkman penalization
2022, Ocean Modelling