Heterogeneities of in- and out-flows in the Mediterranean Sea

摘要

This paper develops a relatively new concept regarding the outflow of Mediterranean Waters (MWs) through the Strait of Gibraltar. While other papers assume that this outflow is composed of only two MWs, we previously found evidence from a re-analysis of 1980s CTD profiles (profiles collected with conductivity-temperature-depth probes during GIBEX, the Gibraltar Experiment), for two other MWs. We also analysed 2003-2008 time series from two CTDs moored (HYDROCHANGES Programme) at the southern sill of Camarinal and on the shelf of Morocco, and we developed a new concept. East of the Strait, the four MWs roughly lay one above the other, but while progressing westward, the associated isopycnals tilt up southward. In the Strait, the MWs are thus juxtaposed, and they all mix with one of two Atlantic Water components (the inflow acronym is thus AWs), and the outflow is horizontally heterogeneous. West of the Strait, the outflow progressively becomes vertically heterogeneous again, hence splitting into a series of superimposed veins. We compared these CTD time series with one collected at the southern sill of Espartel (University of Malaga, INGRES projects and HYDROCHANGES Programme). Fortunately, the CTDs moored at the two sills were generally along the same streamlines so that the MWs' evolution could be monitored. We demonstrated the significance of mixing lines computed from two successive records and the possibility of linking two sets of data (such as CTD profiles) collected at different locations along the Strait, The outflow, which does not show any clear seasonal variability before the Strait, strongly mixes with the inflow within the Strait. This is due mainly to the internal tide and, because the inflow is seasonally variable, leads to an outflow that displays marked seasonal and fortnightly variabilities. Both the outflow and the inflow also display marked spatial heterogeneity and both long-term/yearly and short-term/daily temporal variabilities before they mix; thus, accurately predicting the outflow characteristics in the Atlantic Ocean appears almost impossible.Herein, we first propose a fully objective description of the AWs and MWs during two GIBEX campaigns. Where the AWs and the MWs do not markedly mix, they are defined in terms of density and temperature ranges. Where a MW mixes with one of the AWs down to the bottom, the mixing line characteristics allow for that MW to be followed from one section to one downstream and for the validation of our concept: while superimposed east of the Strait, the MWs come to be juxtaposed within the Strait before becoming superimposed again. We also analysed additional CTD time series collected by the University of Malaga on the south and north sides of the southern sill of Espartel. We demonstrate the following: (a) even though the MWs at the sill (E) and on the south side (ES) were roughly the same, the densest ones out-flowed at ES, i.e., at depths shallower than at E, (b) the MWs on the north side (EN) were very different from those at E and each mixed with different AWs, and (c) using the mixing lines computed from each time series, the data recorded at E and ES allow for the retrieval, with good accuracy, of those recorded at Camarinal (C), which is not the case for the data recorded at EN. Finally we emphasise how different the AWs' heterogeneities are from the MWs' heterogeneities. The inflow is sucked into the Mediterranean Sea, due to the water budget (E-P) deficit there, and it can be composed of any type of AW present west of the Strait at any time and any specific location. The outflow is a product of the Mediterranean Sea, which is like a machine producing a series of MWs that first circulate as alongslope density currents before entering the Strait in a specific order and at specific locations.Consequently, we attempt to schematise the AWs-MWs mixing processes and our understanding of the outflow dynamics. Notwithstanding the difficulty of the working conditions within such a narrow strait, having up to four MWs outflowing side-by-side and mixing with two AWs that have a heterogeneous and variable distribution clearly leads to spatial and temporal heterogeneities that are actually much larger than the ones that have been observed up to now from a relatively low number of CTD profiles and time series.