SATELLITE OBSERVATIONS OF COLD FILAMENTS IN THE MEDITERRANEAN SEA
R. D’Archino
1
, F. Bignami
2
, E. Böhm
2
, E. D’Acunzo
2
, E. Salusti
3*
1
Dip. Biologia Vegetale, Università “La Sapienza”, Rome, Italy
2
ISAC-CNR Sezione di Roma, Rome, Italy
3
INFN, Dip. Fisica, Università “La Sapienza”, Rome, Italy
Abstract
Thermal satellite images relative to the years 1997-2000 are analyzed, in order to infer cold filament and surface jet dynamics in the
Mediterranean Sea. The main zones in which these phenomena are seen to occur are characterised by upwelling and/or the funnelling of
strong cold winds by a somewhat irregular coastal orography. In the Mediterranean Sea the geographical zones with a higher frequency in
these jets are the two lobes of the southern Sicilian coast, the sea off Olbia in Eastern Sardinia, that south of the island of Crete, and the
Balkanic coast of the Adriatic Sea.
Keywords: Filaments, entrainment, potential vorticity, Mediterranean Sea
Rapp. Comm. int. Mer Médit., 37,2004
91
The problem
Thermal and ocean color satellite imagery has recently allowed
considerable progress in the observation and modelling of the
response of coastal currents to strong air-sea interaction. A
particularly interesting case is that of transient jets or cold filaments
observed off coastal areas. For a detailed overview, see [1] and, more
recently, [2]. In all these studies, filament space scales are found to be
100-300 km in length, 10-20 km in width, and 30-50 m in thickness;
the time scales are of one-two weeks, and typical velocities are 10-50
cm s
-1
, irrespective of the elementary physical mechanism originating
the phenomenon. This regularity of filament characteristics,
independently from the physical mechanisms at its origin, is an open
and interesting physical problem. For all the above reasons, a set of
thermal satellite images of cold filaments or jets in the Mediterranean
Sea was examined .These images allowed to estimate the entrainment
coefficient E
*
from the analysis of along-?ow warming.
The images
The entire set of NOAA AVHRR SST images relative to the period
1997-2000 was acquired, processed, archived and analyzed using the
DSP software at the ISAC-CNR Sezione di Roma. The number of
examined images is about 1500 for each year (4-5 daily passes,
relative to the NOAA 14 satellite). The attention was focused on the
southern Sicilian coast, the eastern Sardinian coast, the sea south of
the island of Crete, and also the Balkanic coast of the Adriatic sea as
these sites were found to be the most abundant in filament events.
A set of 44 images distributed between these relevant sites was
finally chosen for a quantitative analysis. For each analyzed cold
filament, at selected cross sections characterized by progressive
numbers, the filament width W(
?
)and the difference between the
environment and filament mean temperatures
?
T(
?
)have been
measured,
?
being the along-stream coordinate. Furthermore, the total
along-stream length L, mean temperature T
m
and meanwidth W
m
have been measured, as well as the average along-stream temperature
difference decrease A =
?
[
?
T(
?
)/
?
T(
?
=0)]/
??
between two given
sections, hereafter referred to as ‘thermal slope’. Time series of SST
and chlorophyll images reveal that during their fully developed phase
one can reasonably assume filament stationariety on a time scale of
one-two days.
The analysis of only a few cross sections per filament scene has
thus induced us to calculate a more reliable estimation of A, using the
entire collection of filaments at each site, with a linear regression
between
?
and
?
T(
?
)/
?
T(
?
=0). We assume that this along-stream
thermal variability of the filaments is due to entrainment of warmer
ambient water and proceed to estimate an ‘overall’entrainment
coefficient E
*
from the thermal images, after establishing the
relationship between A and E
*
. We adopt here the definition of E
*
given by [3]), which assumes that the increment in filament thickness
can be approximated by E
*
|u
|, i.e. proportional to the filament’s mean
along-stream velocity modulus |u
|.
In synthesis, it is interesting to note that the order of magnitude of
the overall E
*
/hvalues stays the same, which suggests this value to be
typical for Mediterranean Sea. The data show some low correlation
coefficients indicative of large scatter in the regressions, as in the case
of the north Adriatic (R=0.30) and eastern Crete (R=-0.13), which
cannot be regarded as statistically significant. Finally, the use of
satellite SST as ambient water temperature is among the limits of
validity of this analysis, sinceT
a
should be a lower ‘bulk’average
temperature. This is because entrainment occurs along the entire
filament interface, which shoals only towards the edge of the filament;
the same holds for h, which therefore is considered as a mean depth
in our conceptual model.
It could also be of interest that during the SYMPLEX 99 cruise (21
October – 6 November 1999, carried out by ISAC-CNR Sezione di
Roma) a CTD transect allowed to determine the thermal structure of
the cold filament ?owing off Cape Passero, at the southernmost tip of
Sicily. The CTD transect is
˜
25 km long; the thermal anomaly is
?
T
˜
2°C, the potential density
s
vertical section reveals that the filament
is moving southward with a velocity
˜
10-20 cm s
-1
. In reality, this
velocity estimate is a lower limit, expressing the bottom to surface
shear. Indeed ADCP measurements during the cruise revealed
absolute velocities up to 50 cm/s. A fact of general interest is that the
filament is in contact with underlying colder waters, while only in the
upper 20 m, say, it comes in contact with warmer waters. This gives
an idea of our approximation in treating the filament as embedded in
warmer homogeneous water.
References
1-Barton, E.D., M.L. Argote, J. Brown, P.M. Kosro, M. Lavin, J.M.
Robles, R.L. Smith, A. Trasvina, and H.S. Velez, 1993. Supersquirt:
Dynamics of the Gulf of Tehuantepec, Mexico. Oceanography, 6, 1: 23-
30.
2-Rienecker, M.M., and C.N.K. Mooers, 1989. Mesoscale eddies, jets
and front off Point Arena, California July 1986. J. Geophys. Res., 94:
12,555-12,569.
3-Turner, J. S., 1986. Turbulent entrainment: the development of the
entrainment asumption, and its application to geophysical ?ows. J. Fluid
Mech., 173: 431-471.
