Results of investigations in the Sicily Channel (1986-1990)
M. M O R E T r I , * t E. SANSONE,* G. SPEZIE* and A. DE MAIO*
(Received 6 February 1992; in revised form 7 December 1992; accepted 9 June 1993)
Abstract--In the framework of the programme POEM (Physical Oceanography of the Eastern Mediterranean), the Italian area of the Sicily Channel was intensively surveyed from 1986 to 1990. A jet-like meandering flow of Atlantic Water was found at the boundary of a sharp thermal front which changed from one survey to another and is evident in satellite imagery only during the warm season. No particularly energetic area was found in the Levantine Intermediate Water (LIW), which is traced by the position of its salinity maximum. A section north of the Pantelleria Island, crossing the deepest part of the Strait, was considered for computing the AW and LIW transports as well as their seasonal changes. The new data give transports smaller than 1 × 106 m 3 s 1, somewhat less than reported in the literature. Seasonal variations indicate that the mean transport is smaller in winter than it is in summer.
I N T R O D U C T I O N
Ix is well known that the evaporation excess over precipitation and runoff make the Eastern Mediterranean Sea a concentration basin with respect both to the Western Mediterranean and to the Atlantic Ocean. This produces the two-layer vertical cell with Atlantic Water (AW) at the surface, and Levantine Intermediate Water (LIW) below it, down to about 500 m, observed in the Mediterranean (NIELSEN, 1912; W~)ST, 1961). The mass and salt exchanges through the Straits of Gibraltar and of Sicily balance the budgets of the entire Mediterranean and of the eastern basin respectively.
The dynamics controlling the transports in the Strait of Sicily is very complex because of the particular bottom topography. The large Sicily and Tunisia shelves make the area available to surface AW (from Sicily to Cape Bon) very wide in comparison to its relatively narrow deeper part (between Sicily and the Tunisia shelf) where LIW enters into the Western basin through two sills 365 and 430 m deep (FRASSETrO, 1965). Several projects investigated the Strait physical conditions (DOIN6 et al. , 1976; GARZOLI and MAILLARD,
1979 ; GARRETI', 1983; GRANCINI and MICHELATO, 1987 ; MOLCARD, 1972 ; BETHOUX, 1979; MORELH et al. , 1975; Mosetti and GELSI, 1984). Recent direct current measurements (GRANONI and MICHELATO, 1987; MANZELLA et al . , 1988) from several moorings showed velocities ranging up to 50 cm s -t with events lasting from 2 to 10 days (low frequency). Of course, tidal phenomena and the inertial component (high frequency) are also very important. Owing to their high variability long time series are needed to evaluate the
*Istituto Universitario Navale, Department of Meteorology and Oceanography, Via A. Acton, 38---80133 Napoli, Italy.
tAu tho r to whom correspondence should be addressed.
1181
1182 M. MORETTI et al.
POEM 1986-1990 3 g . 5
~ . o o o o .i,~" o [] ;-:- , . . . . . . . . . ' ~
Fig. 1. Italian network of CTD stations. Transports are discussed through the transect Sicily- Pantelleria Island (see arrow). Dots represent the general network 1986-1990: squares are additional stations of October 1989 and March 1990; crosses are the additional stations of March 1990 in the Ionian Sea. The topography is given with contours of 200, 600 and 1000 m. The two
small circles A and B indicate the sills, respectively 365 and 430 m deep.
residual currents, which are directly related to the salt, heat and mass budgets of the basins.
In 1985 the international project P O E M organized a high resolution coordinated survey of the whole Eastern Mediterranean Basin to provide a modern homogeneous data set, the necessary support for a deeper understanding of the phenomenology (MALANOTTE- RIZZOLI and ROBINSON, 1988). In this framework, the area of the Channel of Sicily was revisited with modern equipment and a readjusted network of observations (DE MAIO et
a l . , 1972, 1974, 1986 and 1990). Some extensions of the project to the Tyrrhenian and to the Ionian Seas also were considered with the aim of providing an accurate and, as far as possible, exhaustive survey for the analysis of some still poorly understood issues, such as the A W and LIW flows and their variability, the amount of LIW passing over the two sills, its pathways toward the Tyrrhenian Sea and the Sardinia Channel, the type of forcing of the Tyrrhenian and Ionian circulation related to the Strait of Sicily.
DATA SET AND METHODS
The network of hydrological stations in the Channel of Sicily extended from Sicily and Pantelleria Island to the Western limit of the Italian waters (Fig. 1). The adverse meteorological conditions and the limited time availability of R.V. B a n n o c k of the Italian
Results of investigations in the Sicily Channel (1986-1990) 1183
Research Council occasionally prevented the full coverage of all the stations, but the transect of Pantelleria, which was selected for evaluating the AW and LIW transports, was always carried out completely.
The transects are nearly perpendicular to the axis of the channel to provide the best picture of the water mass "cores". A few stations are located on both sides of the Strait to gather data on the LIW in that particular area. Inhomogeneities in the network grid, ranging from a few to some tens of nautical miles, do not provide the best scheme for a detailed objective analysis of the data. For this first phase of the study we preferred to carry out a subjective analysis of T and S field, constructing O-S diagrams of each station with the aim of assessing spatial correlations of maximum and minimum salinity (useful tracers of both AW and LIW) and of finding satisfactory hypotesis for the fluxes computations.
A Neil Brown Mark III CTD was used for the measurements. It was frequently calibrated using water samples collected during a number of deep casts and analysed with the Guildline Autosai.
Calibrations also were performed in cooperation with the Osservatorio Geofisico Sperimentale, Trieste. For its good stability and accuracy within + 0.003 in salinity, no corrections were applied to the final data. The high standard of these data was confirmed during the workshop of Modena (PINARDI, 1988) convened for intercalibration and pooling of all the POEM data.
During the CTD casts an acoustic current meter N.B. /DRCM was lowered down to 10 m depth, definitely below the ship's hull, to record the relative currents. The ship's drift was stated with Loran C fix (relative fix accuracy + 30 m). Position sampling was recorded every 30 s and its mean value was referred to the time span of current meter records (from 10 to 20 min). Finally the absolute current vectors were computed.
The total errors estimated by comparing this method with others, keeping into account the ship's fix accuracy, amount to _ 2 cm s 1 provided the sea state is below Beaufort scale 4 (MORErn et al., 1987).
To obtain consistently accurate data, no current measurements were performed with a sea state higher than scale 4, mainly because the wind forcing in the Ekman layer would produce currents poorly correlated with the currents evaluated with the dynamic method.
The above method provides direct surface current measurements at the CTD station without the need of additional ship's time. These discrete data over the whole area of investigation allow a more complete picture of the surface currents.
Obviously, aliasing caused by high frequency motions (tides, inertial motions, transient forcing of the wind) produces a noise that is superimposed on the low frequency motions (geostrophic currents); the results will be useful only when the signal-to-noise ratio is good
The field of geostrophic currents is obtained by adding together the baroclinic and the barotropic component for each pair of stations. Several approaches exist giving more or less satisfactory results for the determination of the barotropic component (LANDOLT- BORNSTEIN, 1989); some of them are based upon sophisticated inverse theories (WuNSCH, 1978). Indeed the problem is substantially undetermined, so it is necessary to analyze carefully the local situation.
Considering that the maximum depth of the sills in the Strait of Siciliy is 450 m the simplest hypothesis on its low frequency dynamics may be to consider only two layers of water moving in opposite directions.
The examination of vertical salinity and density distributions of a summer and of a
1184 M. MOR~Trl et al.
winter POEM section (Fig. 2), when the picture is less complicated because the seasonal thermocline is missing, shows that the horizontal density gradient is positive down to about 180 in where the isopycnal surfaces ot = 29.0 [Fig. 2(a)] and 28.8 [Fig. 2(b)] are practically horizontal. Below 180 m the gradient reverses and becomes smaller. It reverses again below 500 m.
As in the upper 100 m, where the minimum of salinity shows the presence of AW, the motion is out of Fig. 2, it is necessary to have a negative slope of the free sea surface (i.e. lower near the Sicily coasts). The decrease of velocities with increasing depth is consistent with the positive density gradient. The current reversal in the area of high salinity (LIW) requires that the level of no motion must be located at a depth where the density gradient is still positive, namely above 180 m. So, qualitatively speaking, we may deduce that the surface of no motion should be found at a depth above 180 m. Furthermore, since the flow reversal separates AW and LIW, it would be reasonable for the level of no motion to correspond to an isohaline surface. Thus, in summary, the reference surface should be located (a) in the transition layer between AW and LIW; and (b) along an isohaline.
The results will be checked a posteriori with those obtained by the direct measurements and with the structures observed in the satellite imagery.
_ ~ , I i I I i •
SIC]LIA
0m
Fig. 2. Transect of Pantelleria (see arrow NE-SW of Fig. 1). Surveys of September 1987 (a) and November 1980 (b). The small vertical double pointed arrows indicate the depth of S(z) point of inflection. The front is located in the position marked by the vertical thick arrow. Full curves represent salinity; broken curves represent st. The x-axis points towards Sicily; y is perpendicular
to the section and out of the page; z points downward.
Results of investigations in the Sicily Channel (1986-1990) 1185
f ~:a !
-"" '-'I g ,. i o ' , / (-. "#o.. .
\ J ) s+ X_ , f TUNisiA / "1"O b- oJ
i k ; / "" ~ 0 2040. _ cm/s
11"F 1 5 " E
Fig. 3. Comparison of direct current measurements at 10 m to the flow inferred from the dynamic method. Vectors represent the observed currents during CTD casts (scale in the right down of this figure). Full curves represent the dynamic topography 10/150 db in m dyn m -3. Broken curves represent the isolines of salinity minimum. There is a general good agreement between the two results in the part of higher current velocities that follow the axis of the Channel. The survey
required 7 operational days (R.V. Bannock cruise, 21-30 September 1987).
RESULTS
The S(z) profile and the O-S diagrams show that if the A W occupies the upper layer 100 m thick, and the signal of its summer salinity minimum is at about 50 m, the 10 db topography is well representat ive of the dynamics of the AW. We analyse first the data collected during the survey of September 1987. On the basis of the previous hypotesis we should assume as reference level an isohaline surface close to S = 37.8 psu; but, because this surface is sloping, we use the 150 db surface, which is horizontal and matches the isohaline quite well. The 150 db reference level also will be used for all the surveys. The errors resulting from this choice are not large enough to change noticeably the geostrophic velocities at 10 db because the baroclinic signal is mostly contained in the upper 100 m layer.
Vectors representing the direct current observations superimposed on the pattern of dynamic height anomaly at 10/150 db (Fig. 3) show a fair agreement in the region of higher velocities that follows the axis of the Sicily Channel. There the high frequency currents, responsible for the current aliasing, are smaller. Both the current pattern and the satellite picture evidence a jet-like structure related to a front located between the boundary of the AW jet and the cold saline water present over the Sicily continental shelf [Fig. 2(a)]. This front is evident on the satellite imagery (SATMER, 1987) and on the temperature distri- bution at 10 m depth (Fig. 4).
The salinity distribution has a nucleus of low salinity, the "core" of the AW. The jet is mostly situated in the A W core, which is well recognized through this low salinity. Such salinity features also emerged from our data of 1969-1970. If we disregard the variations
1186 M. MORETTI et al.
induced by the summer heating, which makes the density of water lower than 27.0, and consider the corresponding winter section [Fig. 2(b)], we notice a rather similar S and ot distribution in the upper 150 m. This is true also for the LIW salinity, which has the same S = 38.78 maximum in summer and winter, but the area occupied by the LIW is larger in summer.
A W and L I W dynamics
Figure 5 shows the results of the surveys of October 1986 (a), March 1987 (b), August 1988 (c), October 1989 (d) and March 1990 (e) analysed following the same criteria. AW flow is clearly shown by the layer of salinity minimum, which is always present along the meandering jet entering from the west and stretching toward the Ionian Sea. Meanders change from one survey to the next, but in summer they are always clearly associated with thermal structures.
In winter, in the absence of the thermocline, the surface temperature field is rather homogeneous and AW jets, if present, are no longer well seen. So one might wonder which variations of the circulation are due to the presence of the seasonal thermocline, and which of them are caused by the more energetic forcing of the wind during the winter season. No clear observations favor either hypothesis but upwelling over the shelf of Sicily is certainly a summer phenomenon related to the presence of the thermocline, and to topographic wind forced waves (PICClONI et al., 1988).
A study of the flow patterns taking into account the bottom topography suggests some locking to topography of the AW jet due to the Sicily shelf. However, approximate computations of absolute vorticity along the jet axis did not provide a satisfactory understanding of this effect.
\~ ~t ( ~ / - - , / (--26.5~ <>
TUN'S,A / 6.5 ~
x , ~ I ~X _~j~ L3 I 36°N SEPT. 1 9 8 7 ~ .27 5-~[ ~ I
11OE - ~ 15°E
Fig. 4. Survey of September 1987; isotherms at 10 m and mean position of the thermal front (SATMER, 18-24 September 1987).
Results of investigations in the Sicily Channel (1986--1990) 1187
(a)
OCT. 1986
/ / f \ z (I
TUNISIA /
o
) ,<.>~_ s-"~ ...._._.j ~ ' - ~ 38"N
~ ' ~ ~'~3-/'5\ ~- \ ~
"" ", - 361 '4 • \
11°E 1 ~ . o . . . 15"E
(b) J
c. / .~ <D> ~,.>" k . _
% t \ I kX~,'- MAR. 1987 "-t111: E "<~.~- 15"E _] Fig. 5.
The AW dynamics deduced from the POEM data do not show substantial changes from one survey to another, but the AW flow is characterized by energetic mesoscale phenomena superimposed on the slow basic current. The total height difference between AW and local waters over the Sicily continental shelf is nearly constant, 0.10 dyn m, during the entire period of the surveys (MoRETT], 1989).
Throughout the year, the LIW has a clear salinity signal. Its maximum indicates very
1188 M. MORETFI et al.
(c)
(d)
TUNISIA
AUG.1988
37.' -.. ",, ?
o 2940 o.,/s 1 i
.qgi 0°%1 °0 : > , i ' '
, ' / / I , /
~ k ~ ° '\',.L SIC I L I ' " A //
-36"N
~ _ ~ L IB"E
' ,," ,; , - 3 7 . 5 - - - - - ~ ~,,~
' 0 o ; .." I
" - ' / k . - - <;. ,, J SlClLIA / , / t l
",%
OCT. 1989 ".. "J " . l " 15"E
Fig. 5. Continued.
well the path of the L I W and, for extent of the area and the salinity values, is a steady feature of the hydrological structure of the Sicily Channel.
In order to examine its dynamics, we now evaluate the absolute topography of the 300 dbar surface found at the level of salinity maximum. It cannot be referred to 150 dbar because, as previously discussed, this surface is a good horizontal approximation of the level of no motion for the A W but, because of the isohaline slope, it would suffer from
Results of investigations in the Sicily Channel (1986--1990) 1189
(e)
m
Fig. 5. Surveys of October 1986 (a), March 1987 (b), August 1988 (c), October 1989 (d) and March 1990 (e). Each of them lasted 8-10 operational days. Full curves represent the dynamic
topography 10/150 db. Broken curves represent the isohalines.
errors of the same order of the dynamic signal of L IW to be represented. Indeed below 200 dbar the horizontal density gradient decreases considerably. So, to represent the LIW flows in the area of the Strait, the dynamic height 300-500 dbar was evaluated.
The 500 db surface lies in fact in a transitional layer between the upper layer of L IW spreading and the intermediate waters centered at ~1000 db, some of which have been shown to be Aegean in origin ( P O E M GROUp, 1992). The geostrophic currents were of a few cm s -1, with a bet ter defined area near Pantelleria.
There are two differences between the A W and the L IW layers. The latter is thicker than the former. Considering the lower boundary at 450 m (that is the depth of the West sill) the L I W typical thickness is 300 m as compared to 150 m of the AW. Second, no jet structure may be seen in the L IW because, below 150 m, L IW fills the whole basin. Of course, it would not be possible to evidence barotropic mesoscale phenomena through the baroclinic 300/500 dbar component .
Transports
Any circulation model of the Eastern Mediterranean Basin needs to take into account the water exchange with the Western Basin occurring through the Sicily Channel (ROBINSON et al., 1991). The literature reports the results of two-layer models based on T and S measurements ; a few of them are based on direct current records. The results span between 0.6 and 2.4 × 10 6 m 3 s - I , with A W flow usually est imated about 10% larger than
1190 M. MORETTI et al.
the LIW. Recent investigations reveal the importance of mesoscale phenomena on the transports (GRANCINI and MICHELATO, 1987). Of course, in this case, difficulties arise to separate the part of the mean transport due to the thermohaline forcing from the part caused by the mesoscale eddy field.
The Italian surveys provided a good data set with cross sections of the Sicily Channel North and South of the sill. They could not extend from coast to coast but were limited to the Italian area where the deeper part of the Channel is found. For position and data completeness, the section from Sicily to Pantelleria was the best suited for computing transports as it comprises the "cores" of both AW and LIW. This is important for the computations because a specific reference level of no motion has to be located for each pair of stations that must satisfy the following conditions: (a) to be located on the inversion point of salinity along the vertical (it is usually met at S = 38 psu); and (b) to allow the best balance of incoming and outgoing waters through the section.
It is well known that the reference level is the critical point of the dynamic method, because small vertical shifts of the estimated point of inflection of S(z ) may produce large variations of the results. So, in order to get reasonable results, one must proceed with attempts for making the total transport minimum and LIW and AW transports well balanced.
No extrapolation of the flow down to the bottom for each pair of stations was computed because the transect is not far from the sill which is shallower than the transect in consideration.
A second reason of indeterminacy derives from the fact that opposite flows through the section might not exactly correspond to only the AW and to the LIW movements. For instance a meander of the AW jet might cross the section twice, or some of the entering surface waters might have their origin in the Tyrrhenian Sea or on the Sicily continental shelf. No attempt to minimize this source of errors was carried out to avoid the introduction of further arbitrary assumptions that would make the new results no longer comparable with those reported in the literature.
Some effect of mesoscale structures on the mean flow may be recognized with the aid of the data recorded by three currentmeter moorings deployed east of Pantelleria at depths of 85-180 m, close enough to our level of no motion (MANZELLA et al. , 1988). These records indicated that especially during the cold season, the currents are characterized by large cyclonic and anticyclonic motions having periods from 3 to 10 days, and speeds from 15 to 45 cm s -~. The low salinity associated with the anticyclonic gyres indicates that at shallower depths the mesoscale phenomena may still determine important variations of the interface between AW and LIW.
Table 1 presents the volume, salt and heat transports through the section from Sicily to Pantelleria resulting from the Italian POEM data.
Consistent variations of volume, salt and heat transports are easily seen. They reflect the results obtained from the analysis of the hydrological data on the changes of salinity and temperature of the water mass "cores", and confirm the thermohaline nature of the AW and LIW circulation.
The data show that the mean transport is about half as large as the values reported in the literature (MOLCARD, 1972; GARZOLI and MAILLARD, 1979; GARRETT, 1983; GRANCINI and MICHELATO, 1987; MOSETTI and GELSI, 1984). Such a result might depend on the geographic position of the section, on its smaller area or perhaps because it is located at the boundary of gyres controlled by the bottom topography. However, it should be
Results of investigations in the Sicily Channel (1986-1990) 1191
Table 1. Volume transports (106 m 3 s-I), Salt (kg s 1 106), Heat (J s -1 1012) across the Pantelleria transect
AW (incoming) LIW (outgoing)
Transport Transport Date Volume Salt Heat Volume Salt Heat
October 1986 +0.32 12.03 0.382 -0.40 15.30 0.470 March 1987 +0.63 23.51 0.738 -0.60 22.95 0.702 September 1987 +0.61 23.13 0.721 -0.63 24.37 0.738 March 1988 +0.33 12.40 0.387 -0.28 10.85 0.327 August 1988 +0.95 35.95 1.118 -0.94 36.17 1.102 October 1989 +0.62 2 3 . 4 2 0.734 -0.79 30.53 0.925 March 1990 +0.26 9.74 0.305 -0.24 9.17 0.281
Year mean +0.53 20.03 0.63 -0.55 21.33 0.65 Summer mean +0.63 23.63 0.74 -0.69 26.61 0.81 Winter mean +0.41 15.22 0.47 -0.37 14.32 0.44
r e m a r k e d tha t d i f fe rences also exist a m o n g the pas t va lues r e p o r t e d in the l i t e ra tu re , ce r ta in ly d e p e n d i n g on the d i f fe ren t pos i t ion of the sec t ions t aken into cons ide ra t i on and on the d a t a sets ava i lab le .
A specific c o m m e n t mus t be m a d e for the d i rec t cu r ren t m e a s u r e m e n t s o b t a i n e d no t ve ry far f rom Pan t e l l e r i a (GRANCINI and MICHELATO, 1987). T h e du ra t i on of r ecords ( a b o u t 1 yea r ) seems sufficient to de l i nea t e the L I W and A W t r anspo r t s even if the n u m b e r of po in ts of m e a s u r e m e n t s is ve ry small . C o m p u t a t i o n s based on these r ecords give resul ts twice to four t imes as large as those of T a b l e 1 (MANZELLA e t a l . , 1988) bu t it is i m p o r t a n t to r e m a r k tha t these e s t ima te s ref lect the p reva i l ing inf luence of mesosca le p h e n o m e n a .
The seasona l va r i a t ion of some values in Tab le 1 (0.6 and 0.2 × 106 m 3 s - 1 in March 1987 and 1990; 0.6 x 106 m 3 s -1 in S e p t e m b e r 1987 and O c t o b e r 1989; 0.9 x 106 m 3 s -1 in
A u g u s t 1988) ind ica tes tha t the win te r m e a n t r a n s p o r t is lower than dur ing the s u m m e r , bu t this p o i n t needs fu r the r inves t iga t ion . MANZELLA et al. (1990) r e p o r t tha t in win te r the A W jet gene ra l ly shifts wes twa rd , t owards the coas ts of Af r i ca so one might expec t it to pass ou t s ide the Pan te l l e r i a t ransec t . In these cond i t ions , ba lanc ing the A W and L I W flows m a y l ead to the a b o v e low values .
A s the annua l va r i ab i l i ty of A W and L I W flow is l inked to the c l imat ic cond i t ions of the E a s t e r n M e d i t e r r a n e a n , we no te tha t the win te r t e m p e r a t u r e s of the years 1987-1990 were s o m e w h a t a b o v e the mean . Thus we may specu la te tha t the low values of t r anspo r t may be due to the weak sur face baroc l in ic forc ing of tha t p e r i o d fol lowing the dec rea sed fo rma t ion of sur face dense wa te r .
REFERENCES
BETHOUX J. P. (1979) Budgets of the Mediterranean Sea. Their dependence on the climatc and on the characteristics of the Atlantic waters. Oceanologica Acta, 2(2), 157-163.
DE MAtO A., M. MORET13 and E. SANSONE (1974) Risultati di ricerche idrologiche compiute nel Canale di Sicilia- Ottobre 1969. Accademia Nazionale dei Lincei, Roma-Anno CCCLXXI Quaderno n. 206.
DE MAIO A., M. MOREJTI, E. SANSONE and D. BREGANT (1972) Idrologia del Canale di Sicilia-Nave Dectra, Ott. 1969 Studi in Onore di G. Aliverti, I.U.N. Napoli.
1192 M. MORETTI et al.
DE MAIO A., M. MORETTI, E. SANSONE and G. SPEZIE (1990) Seasonal salinity variations of the LIW entering into the South Tyrrhenian Sea. Il Nuovo Cimento, Vol. 13C.
DOING W., O. BROWN and R. EVANS (1976) Interim Report: a current survey in the Straits o f Sicily. University of Miami.
FRASSETrO R. (1965) A study of the turbulent flow and character of the water masses over the Sicilian Ridge in both summer and winter. Rapports et Proc~s-Verbaux CIESM, 18,812-815.
GARREVI" C. (1983) Variable sea level and strait flows in the Mediterranean: a theoretical study of the response to meteorological forcing. Oceanologica Acta, 6(1), 79-87.
GARZOLI S. and C. MAILLARD (1979) Winter circulation in the Sicily and Sardinia Straits region. Deep-Sea Research, 26A, 933-954.
GRAN£1NI G. F., A. LAVENIA and F. MOSETTI (1972) A contribution to the hydrology of the Strait of Sicily. Proceedings, SACLANT Conference 7, pp. 68-81.
GRANCINI G. F. and A. MICHELATO (1987) Current structure and variability in the Strait of Sicily and adjacent area. Annales Geophysicae, 5B(1), 75-88.
LANDOLT-BORNSTEIN (1989) Oceanography. Vol. 3/b. Sect. 4.4.6.1. Springer-Verlag, Berlin. MALANOTTE-RIZZOLI P. and A. R. ROBINSON (1988) POEM: Physical Oceanography of the Eastern Mediterra-
nean. EOS, 69, 4. MANZELLA G. M. R., G. P. GASPARINI and M. ASTRALDI (1988) Water exchange between the Eastern and
Western Mediterranean through the Strait of Sicily. Deep-Sea Research, 35(6). MANZELLA G. M. R., T. S. HOPKINS, P. J. MINNET and E. NACINI Atlantic Water in the Strait of Sicily. J.C.R.
Vol. 95 No/C2 February 1990. MOLCARD R. (1972) Preliminary results of current measurements in the Strait o f Sicily in May 1970. Proceedings,
SACLANT Conference 7, pp. 82-95. MORELLI C., C. GANTAR and M. PISANI (1975) Bathometry, gravity and magnetism in the Straits of Sicily and in
the Ionian Sea. Boll. Geol. Teor. e. Appl. XVII, 39-58. MORETn, M., E. SANSONE, G. SPEZIE and M. VULTAGGIO (1974) Distribuzione di 7~C, S, Ste 02 nel Canale di
Sicilia. Nave Dectra 1969-70. id.: Batitermogrammi nel Canale di Sicilia. Annali dell'I.U.N, di Napoli, Vol XLI-XLI1. MORETrI, M. (1975) Sul confine fra acque Atlantiche e Levantine nel Canale di Sicilia. Annali dell'IUN, XLIII-
XLIV, 1975 (Napoli). MORETrl M., E. SANSONE and G. SPEZIE (1987) Current measurements from a drifting ship. Boll. Oeeanologia
Teorica ed Applicata, V(3). MORETrl M. (1989) Sea level variability in the Strait of Sicily. Following Changes in the Marine Currents. Boll.
Oceanologia Teorica ed Applicata, VII(4), Ottobre 1989. MOSEyrl F. and G. GELSI (1984) An attempt to calculate the water flow through the Strait of Sicily. B.O.T.A.,
II(1), Trieste, pp. 53-61. NIELSEN J. (1912) Hydrography of the Mediterranean and adjacent waters. Report on the Danish Oc. Expeditions
1908-1910, Copenhagen pp. 72-191. POEM GROUP (1992) General circulation of the Eastern Mediterranean, ROBINSON A. R., P. MALANO1q'E-
RIZZOL1, A. HECHT et al. Earth Science Reviews, 32,285-309. PICEIONI A., M. GABR1ELE, E. SALUSTI and E. ZAMBIANCH1 (19??) Wind induced upwellings off the southern coast
of Sicily. Oceanologica Acta, 11(4). P1NARDI N. (1988) POEM V--87: General Circulation Survey data set preparation. Modena 7-8 March 1988.
IMGA-CNR. T.R.n. pp. 1-88. ROBINSON A. R., M. GOLNARAGHI, W. G. LESLIE, A. ARTEGIANI, A. HECHT, E. LAZZON1, A. MICtlELATO, E.
SANSONE, A. THEOCRA~IS and U. UNLUATA (1991) The eastern Mediterranean general circulation: features, structure and variability. Dynamics of Atmospheres and Oceans, Elscvier, Amsterdam.
SATME~: Bulletin 49, Sept 1987. Le centre de M6t. Spatiale-Lannion. WUNSCH C. (1978) Review of Geophysics and Space Physics, Vol. 16. WOsT G. (1961) On the vertical circulation of the Meditcrranean Sea. Journal of Geophysical Research, 66,
