148
1. Luterbacher and E. Xop1aki
200Jb). Thus, on a wider scale, changes in the temperature and precipitation
distribution are influenced by changes in the atmospheric circulation, which in turn
alter the storm tracks and the temperature and precipitation distribution.
Over the last five centuries the connection between the mean winter precipitation
over the Mediterranean and the NAO reconstructed by Cook et a1. (2001) turned out
to be stable with highly negative correlations throughout the whole period (Table I).
This might be explained by a similar response of the Mediterranean precipitation
behavior on the NAO phase. These findings are in accordance with van Loon and
Rogers (1978), Lamb and Peppler (1987), Hurrell (1995, 1996), Cullen and de
Menocal (2000) and Xoplaki et a1. (200 I b) who found negative correlations between
the winter NAO and station precipitation over most ofthe Mediterranean area and
its vicinity (except for the most southeastern part of the Mediterranean) for the 20 th
century. Our results suggest, that the close relationship between winter NAO and
winter mean Mediterranean precipitation observed in the 20 th century may have been
robust at least back to AD 1500.
The low correlation between the winter NAO and the Mediterranean winter
temperature time series over the last 496 years (Table 1) might be interpreted in
terms of compensatory effects between the areas with negative correlations with the
NAO (Eastern Mediterranean), areas with non significant correlations (Western and
Central Mediterranean), and areas with positive correlations which has been found
for parts of the 20 th century (Cullen and deMenocal 2000; Xoplaki et al. 2001b).
Xoplaki et a1. (200 1 b) used around 250 station ti me series covering the 50 year
period 1949-1999 and found, that the Eastern Mediterranean (Greece, Turkey,
Libya, Egypt, Syria, Jordan, Cyprus, Israel) temperature is negatively correlated with
the NAO, whereas stations mainly North of the Mediterranean Sea (North of 40 N
and East of S'E) show a positive connection to the NAO. Stations on the Iberian
Peninsula, over Italy and Morocco, Algeria and Tunisia return non-significant
correlations with respect to wintertime temperature. With a reduced data set back
to 1900, the main features have been confirmed.
These results suggest to divide the Mediterranean region in various sub areas
and conduct the analysis again. First results indicate, that the NAO plays an
important role in winter variability and leaves a significant signature in temperature
in the eastern Mediterranean whereas the connection with the western and central
Mediterranean is small over the whole 496 years. It may be related also to not
constant transfer functions over time and the extent to which climatic modes from
the 20 th century where the statistical model is fitted, represents the entire range of
climate variability.
Mann (2001) formulated a yearly NAO index back to the middle of the 18 th
century in terms of large-scale surface temperature fields. He also found, that
interannual temperature variability in the Near and Middle East region (30' -40N;
20~ -50 C E) appears to be closely related to patterns of variation associated with this
temperature based NAO.
However, other patterns seem to be of more importance on multidecadal and
secular time scales. This is in agreement with Felis et a1. (2000) who recently
1. Luterbacher and E. Xop1aki
200Jb). Thus, on a wider scale, changes in the temperature and precipitation
distribution are influenced by changes in the atmospheric circulation, which in turn
alter the storm tracks and the temperature and precipitation distribution.
Over the last five centuries the connection between the mean winter precipitation
over the Mediterranean and the NAO reconstructed by Cook et a1. (2001) turned out
to be stable with highly negative correlations throughout the whole period (Table I).
This might be explained by a similar response of the Mediterranean precipitation
behavior on the NAO phase. These findings are in accordance with van Loon and
Rogers (1978), Lamb and Peppler (1987), Hurrell (1995, 1996), Cullen and de
Menocal (2000) and Xoplaki et a1. (200 I b) who found negative correlations between
the winter NAO and station precipitation over most ofthe Mediterranean area and
its vicinity (except for the most southeastern part of the Mediterranean) for the 20 th
century. Our results suggest, that the close relationship between winter NAO and
winter mean Mediterranean precipitation observed in the 20 th century may have been
robust at least back to AD 1500.
The low correlation between the winter NAO and the Mediterranean winter
temperature time series over the last 496 years (Table 1) might be interpreted in
terms of compensatory effects between the areas with negative correlations with the
NAO (Eastern Mediterranean), areas with non significant correlations (Western and
Central Mediterranean), and areas with positive correlations which has been found
for parts of the 20 th century (Cullen and deMenocal 2000; Xoplaki et al. 2001b).
Xoplaki et a1. (200 1 b) used around 250 station ti me series covering the 50 year
period 1949-1999 and found, that the Eastern Mediterranean (Greece, Turkey,
Libya, Egypt, Syria, Jordan, Cyprus, Israel) temperature is negatively correlated with
the NAO, whereas stations mainly North of the Mediterranean Sea (North of 40 N
and East of S'E) show a positive connection to the NAO. Stations on the Iberian
Peninsula, over Italy and Morocco, Algeria and Tunisia return non-significant
correlations with respect to wintertime temperature. With a reduced data set back
to 1900, the main features have been confirmed.
These results suggest to divide the Mediterranean region in various sub areas
and conduct the analysis again. First results indicate, that the NAO plays an
important role in winter variability and leaves a significant signature in temperature
in the eastern Mediterranean whereas the connection with the western and central
Mediterranean is small over the whole 496 years. It may be related also to not
constant transfer functions over time and the extent to which climatic modes from
the 20 th century where the statistical model is fitted, represents the entire range of
climate variability.
Mann (2001) formulated a yearly NAO index back to the middle of the 18 th
century in terms of large-scale surface temperature fields. He also found, that
interannual temperature variability in the Near and Middle East region (30' -40N;
20~ -50 C E) appears to be closely related to patterns of variation associated with this
temperature based NAO.
However, other patterns seem to be of more importance on multidecadal and
secular time scales. This is in agreement with Felis et a1. (2000) who recently