Climate, Climate Variability and Impacts in the Mediterranean Area: An Overview
37
to the Levantine Basin towards the end of the year. Because the LIW is drawn into
the Aegean and afterwards dumped into the deep sea, the outflow of intermediate
water from the Adriatic Sea and through the Sicilian channel is blocked and reduced
(Wu et al. 2000). A simulation ofthe surface heat flux with ECMWF reanalysis data
indicates a number of cold events over the Aegean Sea also between 1990 and 1994
which would keep this process going.
To explore the involved processes the energy budget at the surface has to be
computed. For the large Mediterranean area this is only possible by using bulk
formulas which are either fed with assimilated operational data respectively are used
in models which are initiated by these data. Assimilated data are data which, with
the aid of models, are gridded into a predefined system so that a dense network of
equally distributed data points is obtained. For model simulations the Mediterranean
version of the Geophysical Fluid Dynamics Laboratory modular ocean model
(MOM) has been used. Neither the state parameters measured by the operational
network nor those produced by models are precisely those one would like to have for
accurate flux determinations of which meteorologists anyhow think that only with
the application of eddy correlation methods one can come close to the real fluxes.
It is therefore not astonishing that between the different approaches discrepancies
exist and one is left with estimates the absolute accuracy of which is not exactly
known.
Castellari et al. (1998) investigated the air-sea interactions with a nine years data
set (1980 - 1988) of the 12 hourly NMC atmospheric analysis combined with SST
data of Reynolds (1988) and cloud coverage of COADS (Comprehensive OceanAtmosphere Data Set, da Silva et al. 1994). The 1981 extreme event caused
extraordinary strong sensible (maximum 80 W m 2 ), as well as latent (maximum 280
W m· 2 ) heat fluxes during the event in winter followed by a year of extremely small
wintery fluxes (maxima 34 respectively 140 W m· 2 ). The Aegean event was followed
by two summers with very low evaporation (minimum 20 W m· 2 instead of about 70
W m· 2 ). The surface temperature and total heat flux fields are correlated with a time
lag of 2 - 3 months with the total heat flux leading the SST field.
Korres et al. (2000a and b) investigated in great spatial detail the ocean response
to low frequency atmospheric forcing for the years J 980 - 1989. Their results with
respect to the surface heat budget and transports are reproduced in Table 2. The
strong interannual variability and the response to the extreme events in 1981 and
1986 can clearly be seen. Their general conclusion, underpinned by detailed results
about wind stress, energy budgets, and kinetic energy, is, that the circulation ofthe
Mediterranean indeed reacts to anomalous atmospheric forcing which occurred in
the years 1981 and 1986. Largest amplitudes occur along the Spanish continental
shelf, the western coasts of Corsica and Sardinia as well as in the central Levantine
and Ionian.
The model simulated a surface heat loss of the eastern Mediterranean of 5.3 W
m· 2 in "normal" years which increased to 7.7 W m o2 in cold years. At the same time
E - P changed from 55 to 58 cm year o
\
for the whole Mediterranean while over the
Aegean the increase was 6% and over the eastern Mediterranean 11 % corresponding
37
to the Levantine Basin towards the end of the year. Because the LIW is drawn into
the Aegean and afterwards dumped into the deep sea, the outflow of intermediate
water from the Adriatic Sea and through the Sicilian channel is blocked and reduced
(Wu et al. 2000). A simulation ofthe surface heat flux with ECMWF reanalysis data
indicates a number of cold events over the Aegean Sea also between 1990 and 1994
which would keep this process going.
To explore the involved processes the energy budget at the surface has to be
computed. For the large Mediterranean area this is only possible by using bulk
formulas which are either fed with assimilated operational data respectively are used
in models which are initiated by these data. Assimilated data are data which, with
the aid of models, are gridded into a predefined system so that a dense network of
equally distributed data points is obtained. For model simulations the Mediterranean
version of the Geophysical Fluid Dynamics Laboratory modular ocean model
(MOM) has been used. Neither the state parameters measured by the operational
network nor those produced by models are precisely those one would like to have for
accurate flux determinations of which meteorologists anyhow think that only with
the application of eddy correlation methods one can come close to the real fluxes.
It is therefore not astonishing that between the different approaches discrepancies
exist and one is left with estimates the absolute accuracy of which is not exactly
known.
Castellari et al. (1998) investigated the air-sea interactions with a nine years data
set (1980 - 1988) of the 12 hourly NMC atmospheric analysis combined with SST
data of Reynolds (1988) and cloud coverage of COADS (Comprehensive OceanAtmosphere Data Set, da Silva et al. 1994). The 1981 extreme event caused
extraordinary strong sensible (maximum 80 W m 2 ), as well as latent (maximum 280
W m· 2 ) heat fluxes during the event in winter followed by a year of extremely small
wintery fluxes (maxima 34 respectively 140 W m· 2 ). The Aegean event was followed
by two summers with very low evaporation (minimum 20 W m· 2 instead of about 70
W m· 2 ). The surface temperature and total heat flux fields are correlated with a time
lag of 2 - 3 months with the total heat flux leading the SST field.
Korres et al. (2000a and b) investigated in great spatial detail the ocean response
to low frequency atmospheric forcing for the years J 980 - 1989. Their results with
respect to the surface heat budget and transports are reproduced in Table 2. The
strong interannual variability and the response to the extreme events in 1981 and
1986 can clearly be seen. Their general conclusion, underpinned by detailed results
about wind stress, energy budgets, and kinetic energy, is, that the circulation ofthe
Mediterranean indeed reacts to anomalous atmospheric forcing which occurred in
the years 1981 and 1986. Largest amplitudes occur along the Spanish continental
shelf, the western coasts of Corsica and Sardinia as well as in the central Levantine
and Ionian.
The model simulated a surface heat loss of the eastern Mediterranean of 5.3 W
m· 2 in "normal" years which increased to 7.7 W m o2 in cold years. At the same time
E - P changed from 55 to 58 cm year o
\
for the whole Mediterranean while over the
Aegean the increase was 6% and over the eastern Mediterranean 11 % corresponding