Climate, Climate Variability and Impacts in the Mediterranean Area: An Overview
61
the results derived from their application are only trustworthy if the underlying
climate scenarios can be validated, This now has to be investigated,
6.4 Will the Mediterranean Water Cycle Change?
There are three ways to develop scenarios for the future. One is a simple
extrapolation of observed regional trends which can completely be in error if the
large scale circulation changes. The second way is a mixture of model simulation
and empirical knowledge: If it can be demonstrated that precipitation correlates well
to some large scale phenomenon which can be simulated with some reliability the
future precipitation can be deduced from the development of these large scale
phenomenon. The third way is to rely exclusively on climate model scenarios and
to scale down the results of these computations at certain time slices into the
regional scale. This would require a validated scaling method. The application of all
three methods bears risks in the case ofthe Mediterranean area. Small shifts in the
large scale flows respectively circulations not only brings an area under different
climatic regimes but in addition generates changes in the interaction between
motions at different scales down to land-sea circulation and mountain induced
circulation systems which can destroy the present correlations. We shall inspect now
to which conclusions some of the applied methods lead. Because water is the "life
blood" of the Mediterranean we shall specifically address this parameter.
A simple extrapolation of the meteorological records to obtain an indication how
the future may look like does not seem to have great merits in view ofthe complexity
of the interactions with the NAO and the Hadley circulation. As an example, LopezBermudez et al. (1999) deduced from the 130 years precipitation record of Murcia
(Spain) a downward trend from 368 mm in 1860 to 241 in 1996. If one looks closely
at the record this trend is strongly biassed by a very wet period from 1882 to 1895,
when up to 800 mm were recorded within one year. Taking the time period 19101996 rather an upward trend would result due to maxima around 1950, 1973, and
1985. Generally one can conclude that since 1940 the situation in the average is
pretty stable with a large interannual variability.
North African data series (WMO, 1999) show positive as well as negative
temperature trends but also the time series are of different length (Table 6).
Precipitation statistics has to fight with the large variability of the annual rainfall.
In dry areas with a few mm average precipitation per month the standard deviation
is of the order of the average precipitation. At 29' N, 16 E the temperature trend
was 0.015 'Cia (1955 - 1989) , the average rainfall 2.56 mm/month, the standard
deviation ±2.35 mmlmonth and the trend - 0.013 mmJa. At 27.0 N, 14.4 E the
mean precipitation lies at 0.79 mm/month, the standard deviation is ± 0.93 (1930 -
1988). But also the spatial variability in areas of moderate precipitation can be large.
Algiers had a positive precipitation trend of 0.1 mm/a (1841 - 1972) biassed by the
years 1950 - 1970 while Or an (as well as Constantine) had a negative one of the
same magnitude (1841 - 1990) which is strongly biassed by the wet years around
1870. The temperature is very stable at 17.5 'C. Alexandria and Port Said have a
61
the results derived from their application are only trustworthy if the underlying
climate scenarios can be validated, This now has to be investigated,
6.4 Will the Mediterranean Water Cycle Change?
There are three ways to develop scenarios for the future. One is a simple
extrapolation of observed regional trends which can completely be in error if the
large scale circulation changes. The second way is a mixture of model simulation
and empirical knowledge: If it can be demonstrated that precipitation correlates well
to some large scale phenomenon which can be simulated with some reliability the
future precipitation can be deduced from the development of these large scale
phenomenon. The third way is to rely exclusively on climate model scenarios and
to scale down the results of these computations at certain time slices into the
regional scale. This would require a validated scaling method. The application of all
three methods bears risks in the case ofthe Mediterranean area. Small shifts in the
large scale flows respectively circulations not only brings an area under different
climatic regimes but in addition generates changes in the interaction between
motions at different scales down to land-sea circulation and mountain induced
circulation systems which can destroy the present correlations. We shall inspect now
to which conclusions some of the applied methods lead. Because water is the "life
blood" of the Mediterranean we shall specifically address this parameter.
A simple extrapolation of the meteorological records to obtain an indication how
the future may look like does not seem to have great merits in view ofthe complexity
of the interactions with the NAO and the Hadley circulation. As an example, LopezBermudez et al. (1999) deduced from the 130 years precipitation record of Murcia
(Spain) a downward trend from 368 mm in 1860 to 241 in 1996. If one looks closely
at the record this trend is strongly biassed by a very wet period from 1882 to 1895,
when up to 800 mm were recorded within one year. Taking the time period 19101996 rather an upward trend would result due to maxima around 1950, 1973, and
1985. Generally one can conclude that since 1940 the situation in the average is
pretty stable with a large interannual variability.
North African data series (WMO, 1999) show positive as well as negative
temperature trends but also the time series are of different length (Table 6).
Precipitation statistics has to fight with the large variability of the annual rainfall.
In dry areas with a few mm average precipitation per month the standard deviation
is of the order of the average precipitation. At 29' N, 16 E the temperature trend
was 0.015 'Cia (1955 - 1989) , the average rainfall 2.56 mm/month, the standard
deviation ±2.35 mmlmonth and the trend - 0.013 mmJa. At 27.0 N, 14.4 E the
mean precipitation lies at 0.79 mm/month, the standard deviation is ± 0.93 (1930 -
1988). But also the spatial variability in areas of moderate precipitation can be large.
Algiers had a positive precipitation trend of 0.1 mm/a (1841 - 1972) biassed by the
years 1950 - 1970 while Or an (as well as Constantine) had a negative one of the
same magnitude (1841 - 1990) which is strongly biassed by the wet years around
1870. The temperature is very stable at 17.5 'C. Alexandria and Port Said have a