Analysis of Mediterranean Climate Data: Measured and Modelled
127
the calculation of standardized anomaly indices (SAls) IS:
1 i=n
SAl = -;; L [(Xij - xi )ai ]
i=1
(1)
where n is the number of stations, x is the precipitation at station i for period j, i is
the mean and a is the standard deviation.
Composite precipitation series have been produced for four regions in the
Mediterranean. Some selected results are shown in Figure 1. What is immediately
apparent is that, in general, the trend is towards decreasing precipitation, with a
turning point in the early 1960s. Although these graphs have been selected to show
the trend most clearly, inspection of the full results shows a decreasing trend for
almost all regions in all seasons. The only clear positive trend is in the eastern
Mediterranean in autumn.
These graphs extend to the early 1990s. Extension of the time series to the end
of the 1990s, using the NCEP re-analyses, shows that these downward trends appear
to level off.
3 Understanding the Underlying Causes of the Observed
Trends
The SAls were correlated with two measures of spatial sea level pressure (SLP)
difference. The first is the North Atlantic Oscillation (NAO), calculated from the
pressure difference between Ponta Delgada in the Azores and Stykkisholmur in
Iceland. It is an indicator of activity in the North Atlantic. The second measure is
the difference between SLP at Gibraltar (Northern Frontier) and Israel (Lod
Airport). The existence of a Mediterranean pressure oscillation has been suggested
by a number of authors, based on the similarity between the east-west distance across
the basin and the length of atmospheric long waves. Some results from the
correlation analyses are shown in Table 1.
In the winter season, the NAO is significantly correlated with the western
Mediterranean SAl, as predicted in the literature. However, the relationship appears
to be more extensive than indicated, since significant relationships are also found
with the northern and eastern Mediterranean SAls. The MPI is significantly
correlated with a number of the SAls in winter, but since the sign is the same as that
for the NAO, we cannot assume at this stage that any additional variance is
explained - this may simply be a function of a strong relationship between the NAO
and the MPI. However, significant relationships are found between the MPI and the
western and northern Mediterranean SAls in autumn, and the central and northern
Mediterranean SAls in summer, when the NAO/SAI relationships are weak.
Further analyses were performed to investigate the validity of the MPI. It is
necessary to show (a) that the MPI is a true indicator of a pressure oscillation, rather
127
the calculation of standardized anomaly indices (SAls) IS:
1 i=n
SAl = -;; L [(Xij - xi )ai ]
i=1
(1)
where n is the number of stations, x is the precipitation at station i for period j, i is
the mean and a is the standard deviation.
Composite precipitation series have been produced for four regions in the
Mediterranean. Some selected results are shown in Figure 1. What is immediately
apparent is that, in general, the trend is towards decreasing precipitation, with a
turning point in the early 1960s. Although these graphs have been selected to show
the trend most clearly, inspection of the full results shows a decreasing trend for
almost all regions in all seasons. The only clear positive trend is in the eastern
Mediterranean in autumn.
These graphs extend to the early 1990s. Extension of the time series to the end
of the 1990s, using the NCEP re-analyses, shows that these downward trends appear
to level off.
3 Understanding the Underlying Causes of the Observed
Trends
The SAls were correlated with two measures of spatial sea level pressure (SLP)
difference. The first is the North Atlantic Oscillation (NAO), calculated from the
pressure difference between Ponta Delgada in the Azores and Stykkisholmur in
Iceland. It is an indicator of activity in the North Atlantic. The second measure is
the difference between SLP at Gibraltar (Northern Frontier) and Israel (Lod
Airport). The existence of a Mediterranean pressure oscillation has been suggested
by a number of authors, based on the similarity between the east-west distance across
the basin and the length of atmospheric long waves. Some results from the
correlation analyses are shown in Table 1.
In the winter season, the NAO is significantly correlated with the western
Mediterranean SAl, as predicted in the literature. However, the relationship appears
to be more extensive than indicated, since significant relationships are also found
with the northern and eastern Mediterranean SAls. The MPI is significantly
correlated with a number of the SAls in winter, but since the sign is the same as that
for the NAO, we cannot assume at this stage that any additional variance is
explained - this may simply be a function of a strong relationship between the NAO
and the MPI. However, significant relationships are found between the MPI and the
western and northern Mediterranean SAls in autumn, and the central and northern
Mediterranean SAls in summer, when the NAO/SAI relationships are weak.
Further analyses were performed to investigate the validity of the MPI. It is
necessary to show (a) that the MPI is a true indicator of a pressure oscillation, rather