Water Exchange through the Strait of Gibraltar

Uwe Send and Burkard Baschek

1. Introduction and Overview

During the EU-Project CANIGO (Canary Islands Azores Gibraltar Observations) several moorings with current meters and acoustic instrumentation were deployed at the eastern entrance of the Strait of Gibraltar and were complemented by intensive ship-board observations.

The measurements of the pilot study (Task 4.1.1) improved the understanding of the complex system of flow through the Strait and helped to design a suitable sampling strategy for the intensive experiment (Task 4.1.4).

The objectives of the work presented here are to:

• test the suitability of observing the exchange flow at the eastern entrance of the Strait;
• compare different methods of observing the transport, including acoustic transmissions.
• explore the horizontal and vertical scales and structures of the flow in order to:
  • improve transport estimates.
    
  • observe long-term variability.
    
  • obtain information on hydraulic control, maximal exchange, and overmixing of the Mediterranean Sea.
    
  • provide ``boundary conditions'' (constraints) for models.
    

Reliable transport observations in the strait represent a serious challenge due to various sampling problems, because:

• tidal variability is dominant, thus all ship observations are non-synoptic;
• much of the exchange transport arises from tidal correlations of currents with interface depth (up to 50% at the Camarinal Sill);
• the spatial structure of the flows is complex.
  
  

The present measurement approach therefore concentrates on the outflow at the eastern entrance of the Strait, because:

• the outflow transport there is unaffected by entrainment;
• due to a smaller amplitude of the vertical movement of the interface and a greater depth, transports are less sensitive to tidal interface variability than at the sill (order of 5%).
  
  

The observations in the eastern part of the strait consisted of (Figure 1.1):

• non-synoptic CTD/lADCP-sections;
• selected time-series of CTD/lADCP-stations over a tidal-cycle;
• rapid quasi-synoptic ship-ADCP sections repeated over a tidal-cycle;
• 3 current meter moorings across the strait (collaboration with J.Lafuente and J.Candela / J.Rico) were deployed during the pilot phase of CANIGO and 7 current meter moorings were deployed during the intensive phase of the experiment (Figure 1.1);
• high and low frequency acoustic cross-strait transmission (collaboration with P.Worcester, SIO).
  
  

All of the ship-board observations and much of the mooring and acoustic work took place during the RV ''Poseidon'' cruises 217 and 234 of IfM Kiel in April 1996 and October 1997.

	Figure 1.1: Mooring and station map. Current meter moorings UN, US (IfM Kiel), S,N,C (U. Malaga), J (WHOI / IHM), I1, I2 (SOC) of the intensive phase of CANIGO. Acoustic instrumentation HLF (IfM Kiel), T1-T3, VLA T1-T2 (SIO) (white lines, pilot phase) . Vessel mounted ADCP sections along various transects during Poseidon 217 and Poseidon 234.

2. Ship-board observations

2.1 Currents and hydrographic data

Rapid vessel-mounted ADCP sections were carried out over a complete M2 tidal cycle (12.5 h) at the eastern entrance of the Strait, at the Camarinal Sill, and west of the Sill (Figure 1.1) in order to average the flow over a tidal cycle. All sections reveal significant vertical and horizontal shear and a spatial structure, which is difficult to resolve with mooring arrays.

	Figure 2.1: ADCP sections [cm/s] at the eastern entrance of the Strait of Gibraltar during different parts of a M2 tidal cycle (Poseidon 234). The time difference between the single sections is ~2 h.

	Figure 2.2: Mean current speed [cm/s] through the eastern entrance of the Strait measured withADCP (cruise p217, p234).		Figure 2.3: Mean current speed [cm/s] through the Strait at the Camarinal Sill measured with ADCP (p234).
	Figure 2.4: Mean current speed [cm/s] through the western section measured with ADCP (p234).		Figure 2.5: Mean temperature [c] at the western section from XBT measurements (p234).

2.2 Internal bore

The internal bore is released at the Camarinal Sill when the outflowing tide weakens and the water is pushed back into the Strait of Gibraltar. It was observed at several locations in the Strait in order to study its evolution. The measurements were carried out with continously repeated CTD casts ("CTD-yoyo") in the upper 300 m and ADCP.

The first measurement (Figure 2.6) was carried out close to the Sill and the second measurement (Figure 2.7) at the eastern entrance of the Strait. The figures show the weakening of the strong signals of the bore in interface depth, current speed, and through the dispersion of the wave package.

	Figure 2.6: Observation of an internal bore (P234) at the Camarinal Sill. The time is relativ to high water in Tarifa. a) Along strait component of the currrent speed [cm/s] (ADCP, coloured) and isohalines (CTD, contour-lines). b) vertical velocitiy [cm/s] and isohalines.

	Figure 2.7: Observation of an internal bore (p234) at the eastern entrance of the Strait. a) Along strait component of the current speed [cm/s] (ADCP, coloured) and isohalines (CTD, contour-lines). b) vertical current speed [cm/s] and isohalines.

2.3 Froude numbers

Ideally, the Strait of Gibraltar can be described as a two-layer system. According to the theory of hydraulic control [L. Armi and D. Farmer, 1986, D. Farmer and L. Armi, 1986] the flow through the Strait is hydraulically controlled and the exchange maximal when the flow east of the contraction at Tarifa Narrows and west of Camarinal sill is supercritical. There, the composite Froude number has values of G₂>1. Assuming that the flow west of the sill is supercritical most of the time, the condition G₂>1 in the east holds for maximal exchange. The composite Froude number G₂ can be estimated from measurements at the eastern entrance of the Strait G₂= F₁₂+F₂₂, with F_i2= u_i2/(1- d₁/d₂) /g/h_i .

The current speed of the two layers u_i was determined by taking the mean value outside the main shear zone. It was used together with the mean values of the density (d₁=1027.2 kg m^-3 and d₂=1029.1 kg m^-3) and the depth of the interface hi measured with the CTD.

The measurements of the two research cruises Poseidon 217 and Poseidon 234 can be used for estimating seasonal variations and the influence of the internal bore on the flow through the Strait:

• the estimates of G₂ are significantly smaller in October 1997 than in April 1996 (Figure 2.8, Figure 2.9 and Figure 2.10). No dependence on the spring neap tidal cycle could not be observed.
• The depth of the interface and the layer of zero velocity is significantly shallower in April 1996 than in October 1997.
  
• the along-strait section of the mean current speed in April 1996 shows a slope of the layer of zero-verlocity being qualitatively consistient with hydraulically controlled flow (Figure 2.11). But this is not the case in October 1997 (Figure 2.12).
  

All these qualitative considerations give indications that the flow is hydraulically controlled in October 1997, but possibly not in April 1996. These observations confirm the study from C. Garrett et. al., [1990].

	Figure 2.8: The composite Froude number G2 (above) for the 37.4-isohaline and for the 38.0-isohaline from a time-station with CTD-jojo and vmADCP at the mooring J during spring tide in spring 1996 (Poseidon 217). The lower figure shows the location of the measurements within a spring neap tidal cycle of the currents in 50 m depth.		Figure 2.9: Estimates of the composite Froude number G2 from a time station with CTD and vmADCP at the mooring J during neap tide in spring 1996 (Poseidon 217) for the 37.4- and 38.0-isohaline.
	Figure 2.10: Estimates of the composite Froude number G2 from a time station with CTD and vmADCP at the mooring J in fall 1997(Poseidon 234) for the 37.4- and 38.0-isohaline.
	Figure 2.11: Along strait section of the u-component of the mean current speed [cm/s] from vmADCP measurements during Poseidon 217.		Figure 2.12: Along strait section of the u-component of the mean current speed [cm/s] from vmADCP measurements during Poseidon 234.

Page 2